зеркало из https://github.com/mozilla/gecko-dev.git
759 строки
22 KiB
C++
759 строки
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/. */
|
|
|
|
#include "SourceBuffer.h"
|
|
|
|
#include <algorithm>
|
|
#include <cmath>
|
|
#include <cstring>
|
|
#include "mozilla/Likely.h"
|
|
#include "nsIInputStream.h"
|
|
#include "MainThreadUtils.h"
|
|
#include "SurfaceCache.h"
|
|
|
|
using std::max;
|
|
using std::min;
|
|
|
|
namespace mozilla {
|
|
namespace image {
|
|
|
|
//////////////////////////////////////////////////////////////////////////////
|
|
// SourceBufferIterator implementation.
|
|
//////////////////////////////////////////////////////////////////////////////
|
|
|
|
SourceBufferIterator::~SourceBufferIterator()
|
|
{
|
|
if (mOwner) {
|
|
mOwner->OnIteratorRelease();
|
|
}
|
|
}
|
|
|
|
SourceBufferIterator&
|
|
SourceBufferIterator::operator=(SourceBufferIterator&& aOther)
|
|
{
|
|
if (mOwner) {
|
|
mOwner->OnIteratorRelease();
|
|
}
|
|
|
|
mOwner = Move(aOther.mOwner);
|
|
mState = aOther.mState;
|
|
mData = aOther.mData;
|
|
mChunkCount = aOther.mChunkCount;
|
|
mByteCount = aOther.mByteCount;
|
|
mRemainderToRead = aOther.mRemainderToRead;
|
|
|
|
return *this;
|
|
}
|
|
|
|
SourceBufferIterator::State
|
|
SourceBufferIterator::AdvanceOrScheduleResume(size_t aRequestedBytes,
|
|
IResumable* aConsumer)
|
|
{
|
|
MOZ_ASSERT(mOwner);
|
|
|
|
if (MOZ_UNLIKELY(!HasMore())) {
|
|
MOZ_ASSERT_UNREACHABLE("Should not advance a completed iterator");
|
|
return COMPLETE;
|
|
}
|
|
|
|
// The range of data [mOffset, mOffset + mNextReadLength) has just been read
|
|
// by the caller (or at least they don't have any interest in it), so consume
|
|
// that data.
|
|
MOZ_ASSERT(mData.mIterating.mNextReadLength <= mData.mIterating.mAvailableLength);
|
|
mData.mIterating.mOffset += mData.mIterating.mNextReadLength;
|
|
mData.mIterating.mAvailableLength -= mData.mIterating.mNextReadLength;
|
|
|
|
// An iterator can have a limit imposed on it to read only a subset of a
|
|
// source buffer. If it is present, we need to mimic the same behaviour as
|
|
// the owning SourceBuffer.
|
|
if (MOZ_UNLIKELY(mRemainderToRead != SIZE_MAX)) {
|
|
MOZ_ASSERT(mData.mIterating.mNextReadLength <= mRemainderToRead);
|
|
mRemainderToRead -= mData.mIterating.mNextReadLength;
|
|
|
|
if (MOZ_UNLIKELY(mRemainderToRead == 0)) {
|
|
mData.mIterating.mNextReadLength = 0;
|
|
SetComplete(NS_OK);
|
|
return COMPLETE;
|
|
}
|
|
|
|
if (MOZ_UNLIKELY(aRequestedBytes > mRemainderToRead)) {
|
|
aRequestedBytes = mRemainderToRead;
|
|
}
|
|
}
|
|
|
|
mData.mIterating.mNextReadLength = 0;
|
|
|
|
if (MOZ_LIKELY(mState == READY)) {
|
|
// If the caller wants zero bytes of data, that's easy enough; we just
|
|
// configured ourselves for a zero-byte read above! In theory we could do
|
|
// this even in the START state, but it's not important for performance and
|
|
// breaking the ability of callers to assert that the pointer returned by
|
|
// Data() is non-null doesn't seem worth it.
|
|
if (aRequestedBytes == 0) {
|
|
MOZ_ASSERT(mData.mIterating.mNextReadLength == 0);
|
|
return READY;
|
|
}
|
|
|
|
// Try to satisfy the request out of our local buffer. This is potentially
|
|
// much faster than requesting data from our owning SourceBuffer because we
|
|
// don't have to take the lock. Note that if we have anything at all in our
|
|
// local buffer, we use it to satisfy the request; @aRequestedBytes is just
|
|
// the *maximum* number of bytes we can return.
|
|
if (mData.mIterating.mAvailableLength > 0) {
|
|
return AdvanceFromLocalBuffer(aRequestedBytes);
|
|
}
|
|
}
|
|
|
|
// Our local buffer is empty, so we'll have to request data from our owning
|
|
// SourceBuffer.
|
|
return mOwner->AdvanceIteratorOrScheduleResume(*this,
|
|
aRequestedBytes,
|
|
aConsumer);
|
|
}
|
|
|
|
bool
|
|
SourceBufferIterator::RemainingBytesIsNoMoreThan(size_t aBytes) const
|
|
{
|
|
MOZ_ASSERT(mOwner);
|
|
return mOwner->RemainingBytesIsNoMoreThan(*this, aBytes);
|
|
}
|
|
|
|
|
|
//////////////////////////////////////////////////////////////////////////////
|
|
// SourceBuffer implementation.
|
|
//////////////////////////////////////////////////////////////////////////////
|
|
|
|
const size_t SourceBuffer::MIN_CHUNK_CAPACITY;
|
|
const size_t SourceBuffer::MAX_CHUNK_CAPACITY;
|
|
|
|
SourceBuffer::SourceBuffer()
|
|
: mMutex("image::SourceBuffer")
|
|
, mConsumerCount(0)
|
|
, mCompacted(false)
|
|
{ }
|
|
|
|
SourceBuffer::~SourceBuffer()
|
|
{
|
|
MOZ_ASSERT(mConsumerCount == 0,
|
|
"SourceBuffer destroyed with active consumers");
|
|
}
|
|
|
|
nsresult
|
|
SourceBuffer::AppendChunk(Maybe<Chunk>&& aChunk)
|
|
{
|
|
mMutex.AssertCurrentThreadOwns();
|
|
|
|
#ifdef DEBUG
|
|
if (mChunks.Length() > 0) {
|
|
NS_WARNING("Appending an extra chunk for SourceBuffer");
|
|
}
|
|
#endif
|
|
|
|
if (MOZ_UNLIKELY(!aChunk)) {
|
|
return NS_ERROR_OUT_OF_MEMORY;
|
|
}
|
|
|
|
if (MOZ_UNLIKELY(aChunk->AllocationFailed())) {
|
|
return NS_ERROR_OUT_OF_MEMORY;
|
|
}
|
|
|
|
if (MOZ_UNLIKELY(!mChunks.AppendElement(Move(*aChunk), fallible))) {
|
|
return NS_ERROR_OUT_OF_MEMORY;
|
|
}
|
|
|
|
return NS_OK;
|
|
}
|
|
|
|
Maybe<SourceBuffer::Chunk>
|
|
SourceBuffer::CreateChunk(size_t aCapacity,
|
|
size_t aExistingCapacity /* = 0 */,
|
|
bool aRoundUp /* = true */)
|
|
{
|
|
if (MOZ_UNLIKELY(aCapacity == 0)) {
|
|
MOZ_ASSERT_UNREACHABLE("Appending a chunk of zero size?");
|
|
return Nothing();
|
|
}
|
|
|
|
// Round up if requested.
|
|
size_t finalCapacity = aRoundUp ? RoundedUpCapacity(aCapacity)
|
|
: aCapacity;
|
|
|
|
// Use the size of the SurfaceCache as an additional heuristic to avoid
|
|
// allocating huge buffers. Generally images do not get smaller when decoded,
|
|
// so if we could store the source data in the SurfaceCache, we assume that
|
|
// there's no way we'll be able to store the decoded version.
|
|
if (MOZ_UNLIKELY(!SurfaceCache::CanHold(finalCapacity + aExistingCapacity))) {
|
|
NS_WARNING("SourceBuffer refused to create chunk too large for SurfaceCache");
|
|
return Nothing();
|
|
}
|
|
|
|
return Some(Chunk(finalCapacity));
|
|
}
|
|
|
|
nsresult
|
|
SourceBuffer::Compact()
|
|
{
|
|
mMutex.AssertCurrentThreadOwns();
|
|
|
|
MOZ_ASSERT(mConsumerCount == 0, "Should have no consumers here");
|
|
MOZ_ASSERT(mWaitingConsumers.Length() == 0, "Shouldn't have waiters");
|
|
MOZ_ASSERT(mStatus, "Should be complete here");
|
|
|
|
// If we've tried to compact once, don't attempt again.
|
|
if (mCompacted) {
|
|
return NS_OK;
|
|
}
|
|
|
|
mCompacted = true;
|
|
|
|
// Compact our waiting consumers list, since we're complete and no future
|
|
// consumer will ever have to wait.
|
|
mWaitingConsumers.Compact();
|
|
|
|
// If we have no chunks, then there's nothing to compact.
|
|
if (mChunks.Length() < 1) {
|
|
return NS_OK;
|
|
}
|
|
|
|
// If we have one chunk, then we can compact if it has excess capacity.
|
|
if (mChunks.Length() == 1 && mChunks[0].Length() == mChunks[0].Capacity()) {
|
|
return NS_OK;
|
|
}
|
|
|
|
// If the last chunk has the maximum capacity, then we know the total size
|
|
// will be quite large and not worth consolidating. We can likely/cheapily
|
|
// trim the last chunk if it is too big however.
|
|
size_t capacity = mChunks.LastElement().Capacity();
|
|
if (capacity == MAX_CHUNK_CAPACITY) {
|
|
size_t lastLength = mChunks.LastElement().Length();
|
|
if (lastLength != capacity) {
|
|
mChunks.LastElement().SetCapacity(lastLength);
|
|
}
|
|
return NS_OK;
|
|
}
|
|
|
|
// We can compact our buffer. Determine the total length.
|
|
size_t length = 0;
|
|
for (uint32_t i = 0 ; i < mChunks.Length() ; ++i) {
|
|
length += mChunks[i].Length();
|
|
}
|
|
|
|
// If our total length is zero (which means ExpectLength() got called, but no
|
|
// data ever actually got written) then just empty our chunk list.
|
|
if (MOZ_UNLIKELY(length == 0)) {
|
|
mChunks.Clear();
|
|
return NS_OK;
|
|
}
|
|
|
|
Chunk& mergeChunk = mChunks[0];
|
|
if (MOZ_UNLIKELY(!mergeChunk.SetCapacity(length))) {
|
|
NS_WARNING("Failed to reallocate chunk for SourceBuffer compacting - OOM?");
|
|
return NS_OK;
|
|
}
|
|
|
|
// Copy our old chunks into the newly reallocated first chunk.
|
|
for (uint32_t i = 1 ; i < mChunks.Length() ; ++i) {
|
|
size_t offset = mergeChunk.Length();
|
|
MOZ_ASSERT(offset < mergeChunk.Capacity());
|
|
MOZ_ASSERT(offset + mChunks[i].Length() <= mergeChunk.Capacity());
|
|
|
|
memcpy(mergeChunk.Data() + offset, mChunks[i].Data(), mChunks[i].Length());
|
|
mergeChunk.AddLength(mChunks[i].Length());
|
|
}
|
|
|
|
MOZ_ASSERT(mergeChunk.Length() == mergeChunk.Capacity(),
|
|
"Compacted chunk has slack space");
|
|
|
|
// Remove the redundant chunks.
|
|
mChunks.RemoveElementsAt(1, mChunks.Length() - 1);
|
|
mChunks.Compact();
|
|
|
|
return NS_OK;
|
|
}
|
|
|
|
/* static */ size_t
|
|
SourceBuffer::RoundedUpCapacity(size_t aCapacity)
|
|
{
|
|
// Protect against overflow.
|
|
if (MOZ_UNLIKELY(SIZE_MAX - aCapacity < MIN_CHUNK_CAPACITY)) {
|
|
return aCapacity;
|
|
}
|
|
|
|
// Round up to the next multiple of MIN_CHUNK_CAPACITY (which should be the
|
|
// size of a page).
|
|
size_t roundedCapacity =
|
|
(aCapacity + MIN_CHUNK_CAPACITY - 1) & ~(MIN_CHUNK_CAPACITY - 1);
|
|
MOZ_ASSERT(roundedCapacity >= aCapacity, "Bad math?");
|
|
MOZ_ASSERT(roundedCapacity - aCapacity < MIN_CHUNK_CAPACITY, "Bad math?");
|
|
|
|
return roundedCapacity;
|
|
}
|
|
|
|
size_t
|
|
SourceBuffer::FibonacciCapacityWithMinimum(size_t aMinCapacity)
|
|
{
|
|
mMutex.AssertCurrentThreadOwns();
|
|
|
|
// We grow the source buffer using a Fibonacci growth rate. It will be capped
|
|
// at MAX_CHUNK_CAPACITY, unless the available data exceeds that.
|
|
|
|
size_t length = mChunks.Length();
|
|
|
|
if (length == 0 || aMinCapacity > MAX_CHUNK_CAPACITY) {
|
|
return aMinCapacity;
|
|
}
|
|
|
|
if (length == 1) {
|
|
return min(max(2 * mChunks[0].Capacity(), aMinCapacity),
|
|
MAX_CHUNK_CAPACITY);
|
|
}
|
|
|
|
return min(max(mChunks[length - 1].Capacity() +
|
|
mChunks[length - 2].Capacity(),
|
|
aMinCapacity), MAX_CHUNK_CAPACITY);
|
|
}
|
|
|
|
void
|
|
SourceBuffer::AddWaitingConsumer(IResumable* aConsumer)
|
|
{
|
|
mMutex.AssertCurrentThreadOwns();
|
|
|
|
MOZ_ASSERT(!mStatus, "Waiting when we're complete?");
|
|
|
|
if (aConsumer) {
|
|
mWaitingConsumers.AppendElement(aConsumer);
|
|
}
|
|
}
|
|
|
|
void
|
|
SourceBuffer::ResumeWaitingConsumers()
|
|
{
|
|
mMutex.AssertCurrentThreadOwns();
|
|
|
|
if (mWaitingConsumers.Length() == 0) {
|
|
return;
|
|
}
|
|
|
|
for (uint32_t i = 0 ; i < mWaitingConsumers.Length() ; ++i) {
|
|
mWaitingConsumers[i]->Resume();
|
|
}
|
|
|
|
mWaitingConsumers.Clear();
|
|
}
|
|
|
|
nsresult
|
|
SourceBuffer::ExpectLength(size_t aExpectedLength)
|
|
{
|
|
MOZ_ASSERT(aExpectedLength > 0, "Zero expected size?");
|
|
|
|
MutexAutoLock lock(mMutex);
|
|
|
|
if (MOZ_UNLIKELY(mStatus)) {
|
|
MOZ_ASSERT_UNREACHABLE("ExpectLength after SourceBuffer is complete");
|
|
return NS_OK;
|
|
}
|
|
|
|
if (MOZ_UNLIKELY(mChunks.Length() > 0)) {
|
|
MOZ_ASSERT_UNREACHABLE("Duplicate or post-Append call to ExpectLength");
|
|
return NS_OK;
|
|
}
|
|
|
|
if (MOZ_UNLIKELY(!SurfaceCache::CanHold(aExpectedLength))) {
|
|
NS_WARNING("SourceBuffer refused to store too large buffer");
|
|
return HandleError(NS_ERROR_INVALID_ARG);
|
|
}
|
|
|
|
size_t length = min(aExpectedLength, MAX_CHUNK_CAPACITY);
|
|
if (MOZ_UNLIKELY(NS_FAILED(AppendChunk(CreateChunk(length,
|
|
/* aExistingCapacity */ 0,
|
|
/* aRoundUp */ false))))) {
|
|
return HandleError(NS_ERROR_OUT_OF_MEMORY);
|
|
}
|
|
|
|
return NS_OK;
|
|
}
|
|
|
|
nsresult
|
|
SourceBuffer::Append(const char* aData, size_t aLength)
|
|
{
|
|
MOZ_ASSERT(aData, "Should have a buffer");
|
|
MOZ_ASSERT(aLength > 0, "Writing a zero-sized chunk");
|
|
|
|
size_t currentChunkCapacity = 0;
|
|
size_t currentChunkLength = 0;
|
|
char* currentChunkData = nullptr;
|
|
size_t currentChunkRemaining = 0;
|
|
size_t forCurrentChunk = 0;
|
|
size_t forNextChunk = 0;
|
|
size_t nextChunkCapacity = 0;
|
|
size_t totalCapacity = 0;
|
|
|
|
{
|
|
MutexAutoLock lock(mMutex);
|
|
|
|
if (MOZ_UNLIKELY(mStatus)) {
|
|
// This SourceBuffer is already complete; ignore further data.
|
|
return NS_ERROR_FAILURE;
|
|
}
|
|
|
|
if (MOZ_UNLIKELY(mChunks.Length() == 0)) {
|
|
if (MOZ_UNLIKELY(NS_FAILED(AppendChunk(CreateChunk(aLength))))) {
|
|
return HandleError(NS_ERROR_OUT_OF_MEMORY);
|
|
}
|
|
}
|
|
|
|
// Copy out the current chunk's information so we can release the lock.
|
|
// Note that this wouldn't be safe if multiple producers were allowed!
|
|
Chunk& currentChunk = mChunks.LastElement();
|
|
currentChunkCapacity = currentChunk.Capacity();
|
|
currentChunkLength = currentChunk.Length();
|
|
currentChunkData = currentChunk.Data();
|
|
|
|
// Partition this data between the current chunk and the next chunk.
|
|
// (Because we always allocate a chunk big enough to fit everything passed
|
|
// to Append, we'll never need more than those two chunks to store
|
|
// everything.)
|
|
currentChunkRemaining = currentChunkCapacity - currentChunkLength;
|
|
forCurrentChunk = min(aLength, currentChunkRemaining);
|
|
forNextChunk = aLength - forCurrentChunk;
|
|
|
|
// If we'll need another chunk, determine what its capacity should be while
|
|
// we still hold the lock.
|
|
nextChunkCapacity = forNextChunk > 0
|
|
? FibonacciCapacityWithMinimum(forNextChunk)
|
|
: 0;
|
|
|
|
for (uint32_t i = 0 ; i < mChunks.Length() ; ++i) {
|
|
totalCapacity += mChunks[i].Capacity();
|
|
}
|
|
}
|
|
|
|
// Write everything we can fit into the current chunk.
|
|
MOZ_ASSERT(currentChunkLength + forCurrentChunk <= currentChunkCapacity);
|
|
memcpy(currentChunkData + currentChunkLength, aData, forCurrentChunk);
|
|
|
|
// If there's something left, create a new chunk and write it there.
|
|
Maybe<Chunk> nextChunk;
|
|
if (forNextChunk > 0) {
|
|
MOZ_ASSERT(nextChunkCapacity >= forNextChunk, "Next chunk too small?");
|
|
nextChunk = CreateChunk(nextChunkCapacity, totalCapacity);
|
|
if (MOZ_LIKELY(nextChunk && !nextChunk->AllocationFailed())) {
|
|
memcpy(nextChunk->Data(), aData + forCurrentChunk, forNextChunk);
|
|
nextChunk->AddLength(forNextChunk);
|
|
}
|
|
}
|
|
|
|
// Update shared data structures.
|
|
{
|
|
MutexAutoLock lock(mMutex);
|
|
|
|
// Update the length of the current chunk.
|
|
Chunk& currentChunk = mChunks.LastElement();
|
|
MOZ_ASSERT(currentChunk.Data() == currentChunkData, "Multiple producers?");
|
|
MOZ_ASSERT(currentChunk.Length() == currentChunkLength,
|
|
"Multiple producers?");
|
|
|
|
currentChunk.AddLength(forCurrentChunk);
|
|
|
|
// If we created a new chunk, add it to the series.
|
|
if (forNextChunk > 0) {
|
|
if (MOZ_UNLIKELY(!nextChunk)) {
|
|
return HandleError(NS_ERROR_OUT_OF_MEMORY);
|
|
}
|
|
|
|
if (MOZ_UNLIKELY(NS_FAILED(AppendChunk(Move(nextChunk))))) {
|
|
return HandleError(NS_ERROR_OUT_OF_MEMORY);
|
|
}
|
|
}
|
|
|
|
// Resume any waiting readers now that there's new data.
|
|
ResumeWaitingConsumers();
|
|
}
|
|
|
|
return NS_OK;
|
|
}
|
|
|
|
static nsresult
|
|
AppendToSourceBuffer(nsIInputStream*,
|
|
void* aClosure,
|
|
const char* aFromRawSegment,
|
|
uint32_t,
|
|
uint32_t aCount,
|
|
uint32_t* aWriteCount)
|
|
{
|
|
SourceBuffer* sourceBuffer = static_cast<SourceBuffer*>(aClosure);
|
|
|
|
// Copy the source data. Unless we hit OOM, we squelch the return value here,
|
|
// because returning an error means that ReadSegments stops reading data, and
|
|
// we want to ensure that we read everything we get. If we hit OOM then we
|
|
// return a failed status to the caller.
|
|
nsresult rv = sourceBuffer->Append(aFromRawSegment, aCount);
|
|
if (rv == NS_ERROR_OUT_OF_MEMORY) {
|
|
return rv;
|
|
}
|
|
|
|
// Report that we wrote everything we got.
|
|
*aWriteCount = aCount;
|
|
|
|
return NS_OK;
|
|
}
|
|
|
|
nsresult
|
|
SourceBuffer::AppendFromInputStream(nsIInputStream* aInputStream,
|
|
uint32_t aCount)
|
|
{
|
|
uint32_t bytesRead;
|
|
nsresult rv = aInputStream->ReadSegments(AppendToSourceBuffer, this,
|
|
aCount, &bytesRead);
|
|
if (NS_WARN_IF(NS_FAILED(rv))) {
|
|
return rv;
|
|
}
|
|
|
|
if (bytesRead == 0) {
|
|
// The loading of the image has been canceled.
|
|
return NS_ERROR_FAILURE;
|
|
}
|
|
|
|
if (bytesRead != aCount) {
|
|
// Only some of the given data was read. We must have failed in
|
|
// SourceBuffer::Append but ReadSegments swallowed the error.
|
|
MutexAutoLock lock(mMutex);
|
|
MOZ_ASSERT(mStatus);
|
|
MOZ_ASSERT(NS_FAILED(*mStatus));
|
|
return *mStatus;
|
|
}
|
|
|
|
return rv;
|
|
}
|
|
|
|
void
|
|
SourceBuffer::Complete(nsresult aStatus)
|
|
{
|
|
MutexAutoLock lock(mMutex);
|
|
|
|
// When an error occurs internally (e.g. due to an OOM), we save the status.
|
|
// This will indirectly trigger a failure higher up and that will call
|
|
// SourceBuffer::Complete. Since it doesn't necessarily know we are already
|
|
// complete, it is safe to ignore.
|
|
if (mStatus && (MOZ_UNLIKELY(NS_SUCCEEDED(*mStatus) ||
|
|
aStatus != NS_IMAGELIB_ERROR_FAILURE))) {
|
|
MOZ_ASSERT_UNREACHABLE("Called Complete more than once");
|
|
return;
|
|
}
|
|
|
|
if (MOZ_UNLIKELY(NS_SUCCEEDED(aStatus) && IsEmpty())) {
|
|
// It's illegal to succeed without writing anything.
|
|
aStatus = NS_ERROR_FAILURE;
|
|
}
|
|
|
|
mStatus = Some(aStatus);
|
|
|
|
// Resume any waiting consumers now that we're complete.
|
|
ResumeWaitingConsumers();
|
|
|
|
// If we still have active consumers, just return.
|
|
if (mConsumerCount > 0) {
|
|
return;
|
|
}
|
|
|
|
// Attempt to compact our buffer down to a single chunk.
|
|
Compact();
|
|
}
|
|
|
|
bool
|
|
SourceBuffer::IsComplete()
|
|
{
|
|
MutexAutoLock lock(mMutex);
|
|
return bool(mStatus);
|
|
}
|
|
|
|
size_t
|
|
SourceBuffer::SizeOfIncludingThisWithComputedFallback(MallocSizeOf
|
|
aMallocSizeOf) const
|
|
{
|
|
MutexAutoLock lock(mMutex);
|
|
|
|
size_t n = aMallocSizeOf(this);
|
|
n += mChunks.ShallowSizeOfExcludingThis(aMallocSizeOf);
|
|
|
|
for (uint32_t i = 0 ; i < mChunks.Length() ; ++i) {
|
|
size_t chunkSize = aMallocSizeOf(mChunks[i].Data());
|
|
|
|
if (chunkSize == 0) {
|
|
// We're on a platform where moz_malloc_size_of always returns 0.
|
|
chunkSize = mChunks[i].Capacity();
|
|
}
|
|
|
|
n += chunkSize;
|
|
}
|
|
|
|
return n;
|
|
}
|
|
|
|
SourceBufferIterator
|
|
SourceBuffer::Iterator(size_t aReadLength)
|
|
{
|
|
{
|
|
MutexAutoLock lock(mMutex);
|
|
mConsumerCount++;
|
|
}
|
|
|
|
return SourceBufferIterator(this, aReadLength);
|
|
}
|
|
|
|
void
|
|
SourceBuffer::OnIteratorRelease()
|
|
{
|
|
MutexAutoLock lock(mMutex);
|
|
|
|
MOZ_ASSERT(mConsumerCount > 0, "Consumer count doesn't add up");
|
|
mConsumerCount--;
|
|
|
|
// If we still have active consumers, or we're not complete yet, then return.
|
|
if (mConsumerCount > 0 || !mStatus) {
|
|
return;
|
|
}
|
|
|
|
// Attempt to compact our buffer down to a single chunk.
|
|
Compact();
|
|
}
|
|
|
|
bool
|
|
SourceBuffer::RemainingBytesIsNoMoreThan(const SourceBufferIterator& aIterator,
|
|
size_t aBytes) const
|
|
{
|
|
MutexAutoLock lock(mMutex);
|
|
|
|
// If we're not complete, we always say no.
|
|
if (!mStatus) {
|
|
return false;
|
|
}
|
|
|
|
// If the iterator's at the end, the answer is trivial.
|
|
if (!aIterator.HasMore()) {
|
|
return true;
|
|
}
|
|
|
|
uint32_t iteratorChunk = aIterator.mData.mIterating.mChunk;
|
|
size_t iteratorOffset = aIterator.mData.mIterating.mOffset;
|
|
size_t iteratorLength = aIterator.mData.mIterating.mAvailableLength;
|
|
|
|
// Include the bytes the iterator is currently pointing to in the limit, so
|
|
// that the current chunk doesn't have to be a special case.
|
|
size_t bytes = aBytes + iteratorOffset + iteratorLength;
|
|
|
|
// Count the length over all of our chunks, starting with the one that the
|
|
// iterator is currently pointing to. (This is O(N), but N is expected to be
|
|
// ~1, so it doesn't seem worth caching the length separately.)
|
|
size_t lengthSoFar = 0;
|
|
for (uint32_t i = iteratorChunk ; i < mChunks.Length() ; ++i) {
|
|
lengthSoFar += mChunks[i].Length();
|
|
if (lengthSoFar > bytes) {
|
|
return false;
|
|
}
|
|
}
|
|
|
|
return true;
|
|
}
|
|
|
|
SourceBufferIterator::State
|
|
SourceBuffer::AdvanceIteratorOrScheduleResume(SourceBufferIterator& aIterator,
|
|
size_t aRequestedBytes,
|
|
IResumable* aConsumer)
|
|
{
|
|
MutexAutoLock lock(mMutex);
|
|
|
|
MOZ_ASSERT(aIterator.HasMore(), "Advancing a completed iterator and "
|
|
"AdvanceOrScheduleResume didn't catch it");
|
|
|
|
if (MOZ_UNLIKELY(mStatus && NS_FAILED(*mStatus))) {
|
|
// This SourceBuffer is complete due to an error; all reads fail.
|
|
return aIterator.SetComplete(*mStatus);
|
|
}
|
|
|
|
if (MOZ_UNLIKELY(mChunks.Length() == 0)) {
|
|
// We haven't gotten an initial chunk yet.
|
|
AddWaitingConsumer(aConsumer);
|
|
return aIterator.SetWaiting(!!aConsumer);
|
|
}
|
|
|
|
uint32_t iteratorChunkIdx = aIterator.mData.mIterating.mChunk;
|
|
MOZ_ASSERT(iteratorChunkIdx < mChunks.Length());
|
|
|
|
const Chunk& currentChunk = mChunks[iteratorChunkIdx];
|
|
size_t iteratorEnd = aIterator.mData.mIterating.mOffset +
|
|
aIterator.mData.mIterating.mAvailableLength;
|
|
MOZ_ASSERT(iteratorEnd <= currentChunk.Length());
|
|
MOZ_ASSERT(iteratorEnd <= currentChunk.Capacity());
|
|
|
|
if (iteratorEnd < currentChunk.Length()) {
|
|
// There's more data in the current chunk.
|
|
return aIterator.SetReady(iteratorChunkIdx, currentChunk.Data(),
|
|
iteratorEnd, currentChunk.Length() - iteratorEnd,
|
|
aRequestedBytes);
|
|
}
|
|
|
|
if (iteratorEnd == currentChunk.Capacity() &&
|
|
!IsLastChunk(iteratorChunkIdx)) {
|
|
// Advance to the next chunk.
|
|
const Chunk& nextChunk = mChunks[iteratorChunkIdx + 1];
|
|
return aIterator.SetReady(iteratorChunkIdx + 1, nextChunk.Data(), 0,
|
|
nextChunk.Length(), aRequestedBytes);
|
|
}
|
|
|
|
MOZ_ASSERT(IsLastChunk(iteratorChunkIdx), "Should've advanced");
|
|
|
|
if (mStatus) {
|
|
// There's no more data and this SourceBuffer completed successfully.
|
|
MOZ_ASSERT(NS_SUCCEEDED(*mStatus), "Handled failures earlier");
|
|
return aIterator.SetComplete(*mStatus);
|
|
}
|
|
|
|
// We're not complete, but there's no more data right now. Arrange to wake up
|
|
// the consumer when we get more data.
|
|
AddWaitingConsumer(aConsumer);
|
|
return aIterator.SetWaiting(!!aConsumer);
|
|
}
|
|
|
|
nsresult
|
|
SourceBuffer::HandleError(nsresult aError)
|
|
{
|
|
MOZ_ASSERT(NS_FAILED(aError), "Should have an error here");
|
|
MOZ_ASSERT(aError == NS_ERROR_OUT_OF_MEMORY ||
|
|
aError == NS_ERROR_INVALID_ARG,
|
|
"Unexpected error; may want to notify waiting readers, which "
|
|
"HandleError currently doesn't do");
|
|
|
|
mMutex.AssertCurrentThreadOwns();
|
|
|
|
NS_WARNING("SourceBuffer encountered an unrecoverable error");
|
|
|
|
// Record the error.
|
|
mStatus = Some(aError);
|
|
|
|
// Drop our references to waiting readers.
|
|
mWaitingConsumers.Clear();
|
|
|
|
return *mStatus;
|
|
}
|
|
|
|
bool
|
|
SourceBuffer::IsEmpty()
|
|
{
|
|
mMutex.AssertCurrentThreadOwns();
|
|
return mChunks.Length() == 0 ||
|
|
mChunks[0].Length() == 0;
|
|
}
|
|
|
|
bool
|
|
SourceBuffer::IsLastChunk(uint32_t aChunk)
|
|
{
|
|
mMutex.AssertCurrentThreadOwns();
|
|
return aChunk + 1 == mChunks.Length();
|
|
}
|
|
|
|
} // namespace image
|
|
} // namespace mozilla
|