blob: 85655d97723634e85c3a7eaacea60d5d3f56545f [file] [log] [blame]
// Copyright (c) 2013 The Chromium Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#include "net/disk_cache/simple/simple_entry_impl.h"
#include <algorithm>
#include <cstring>
#include <limits>
#include <utility>
#include <vector>
#include "base/bind.h"
#include "base/bind_helpers.h"
#include "base/callback.h"
#include "base/location.h"
#include "base/logging.h"
#include "base/single_thread_task_runner.h"
#include "base/task_runner.h"
#include "base/task_runner_util.h"
#include "base/threading/thread_task_runner_handle.h"
#include "base/time/time.h"
#include "net/base/io_buffer.h"
#include "net/base/net_errors.h"
#include "net/disk_cache/net_log_parameters.h"
#include "net/disk_cache/simple/simple_backend_impl.h"
#include "net/disk_cache/simple/simple_histogram_macros.h"
#include "net/disk_cache/simple/simple_index.h"
#include "net/disk_cache/simple/simple_net_log_parameters.h"
#include "net/disk_cache/simple/simple_synchronous_entry.h"
#include "net/disk_cache/simple/simple_util.h"
#include "net/log/net_log_source_type.h"
#include "third_party/zlib/zlib.h"
namespace disk_cache {
namespace {
// An entry can store sparse data taking up to 1 / kMaxSparseDataSizeDivisor of
// the cache.
const int64_t kMaxSparseDataSizeDivisor = 10;
// Used in histograms, please only add entries at the end.
enum ReadResult {
READ_RESULT_SUCCESS = 0,
READ_RESULT_INVALID_ARGUMENT = 1,
READ_RESULT_NONBLOCK_EMPTY_RETURN = 2,
READ_RESULT_BAD_STATE = 3,
READ_RESULT_FAST_EMPTY_RETURN = 4,
READ_RESULT_SYNC_READ_FAILURE = 5,
READ_RESULT_SYNC_CHECKSUM_FAILURE = 6,
READ_RESULT_MAX = 7,
};
// Used in histograms, please only add entries at the end.
enum WriteResult {
WRITE_RESULT_SUCCESS = 0,
WRITE_RESULT_INVALID_ARGUMENT = 1,
WRITE_RESULT_OVER_MAX_SIZE = 2,
WRITE_RESULT_BAD_STATE = 3,
WRITE_RESULT_SYNC_WRITE_FAILURE = 4,
WRITE_RESULT_FAST_EMPTY_RETURN = 5,
WRITE_RESULT_MAX = 6,
};
// Used in histograms, please only add entries at the end.
enum HeaderSizeChange {
HEADER_SIZE_CHANGE_INITIAL,
HEADER_SIZE_CHANGE_SAME,
HEADER_SIZE_CHANGE_INCREASE,
HEADER_SIZE_CHANGE_DECREASE,
HEADER_SIZE_CHANGE_UNEXPECTED_WRITE,
HEADER_SIZE_CHANGE_MAX
};
void RecordReadResult(net::CacheType cache_type, ReadResult result) {
SIMPLE_CACHE_UMA(ENUMERATION,
"ReadResult", cache_type, result, READ_RESULT_MAX);
}
void RecordWriteResult(net::CacheType cache_type, WriteResult result) {
SIMPLE_CACHE_UMA(ENUMERATION,
"WriteResult2", cache_type, result, WRITE_RESULT_MAX);
}
// TODO(juliatuttle): Consider removing this once we have a good handle on
// header size changes.
void RecordHeaderSizeChange(net::CacheType cache_type,
int old_size, int new_size) {
HeaderSizeChange size_change;
SIMPLE_CACHE_UMA(COUNTS_10000, "HeaderSize", cache_type, new_size);
if (old_size == 0) {
size_change = HEADER_SIZE_CHANGE_INITIAL;
} else if (new_size == old_size) {
size_change = HEADER_SIZE_CHANGE_SAME;
} else if (new_size > old_size) {
int delta = new_size - old_size;
SIMPLE_CACHE_UMA(COUNTS_10000,
"HeaderSizeIncreaseAbsolute", cache_type, delta);
SIMPLE_CACHE_UMA(PERCENTAGE,
"HeaderSizeIncreasePercentage", cache_type,
delta * 100 / old_size);
size_change = HEADER_SIZE_CHANGE_INCREASE;
} else { // new_size < old_size
int delta = old_size - new_size;
SIMPLE_CACHE_UMA(COUNTS_10000,
"HeaderSizeDecreaseAbsolute", cache_type, delta);
SIMPLE_CACHE_UMA(PERCENTAGE,
"HeaderSizeDecreasePercentage", cache_type,
delta * 100 / old_size);
size_change = HEADER_SIZE_CHANGE_DECREASE;
}
SIMPLE_CACHE_UMA(ENUMERATION,
"HeaderSizeChange", cache_type,
size_change, HEADER_SIZE_CHANGE_MAX);
}
void RecordUnexpectedStream0Write(net::CacheType cache_type) {
SIMPLE_CACHE_UMA(ENUMERATION,
"HeaderSizeChange", cache_type,
HEADER_SIZE_CHANGE_UNEXPECTED_WRITE, HEADER_SIZE_CHANGE_MAX);
}
int g_open_entry_count = 0;
void AdjustOpenEntryCountBy(net::CacheType cache_type, int offset) {
g_open_entry_count += offset;
SIMPLE_CACHE_UMA(COUNTS_10000,
"GlobalOpenEntryCount", cache_type, g_open_entry_count);
}
void InvokeCallbackIfBackendIsAlive(
const base::WeakPtr<SimpleBackendImpl>& backend,
const net::CompletionCallback& completion_callback,
int result) {
DCHECK(!completion_callback.is_null());
if (!backend.get())
return;
completion_callback.Run(result);
}
} // namespace
using base::Closure;
using base::FilePath;
using base::Time;
using base::TaskRunner;
// A helper class to insure that RunNextOperationIfNeeded() is called when
// exiting the current stack frame.
class SimpleEntryImpl::ScopedOperationRunner {
public:
explicit ScopedOperationRunner(SimpleEntryImpl* entry) : entry_(entry) {
}
~ScopedOperationRunner() {
entry_->RunNextOperationIfNeeded();
}
private:
SimpleEntryImpl* const entry_;
};
SimpleEntryImpl::ActiveEntryProxy::~ActiveEntryProxy() {}
SimpleEntryImpl::SimpleEntryImpl(net::CacheType cache_type,
const FilePath& path,
const uint64_t entry_hash,
OperationsMode operations_mode,
SimpleBackendImpl* backend,
net::NetLog* net_log)
: backend_(backend->AsWeakPtr()),
cache_type_(cache_type),
worker_pool_(backend->worker_pool()),
path_(path),
entry_hash_(entry_hash),
use_optimistic_operations_(operations_mode == OPTIMISTIC_OPERATIONS),
last_used_(Time::Now()),
last_modified_(last_used_),
sparse_data_size_(0),
open_count_(0),
doomed_(false),
state_(STATE_UNINITIALIZED),
synchronous_entry_(NULL),
net_log_(net::BoundNetLog::Make(net_log,
net::NetLogSourceType::DISK_CACHE_ENTRY)),
stream_0_data_(new net::GrowableIOBuffer()) {
static_assert(arraysize(data_size_) == arraysize(crc32s_end_offset_),
"arrays should be the same size");
static_assert(arraysize(data_size_) == arraysize(crc32s_),
"arrays should be the same size");
static_assert(arraysize(data_size_) == arraysize(have_written_),
"arrays should be the same size");
static_assert(arraysize(data_size_) == arraysize(crc_check_state_),
"arrays should be the same size");
MakeUninitialized();
net_log_.BeginEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY,
CreateNetLogSimpleEntryConstructionCallback(this));
}
void SimpleEntryImpl::SetActiveEntryProxy(
std::unique_ptr<ActiveEntryProxy> active_entry_proxy) {
DCHECK(!active_entry_proxy_);
active_entry_proxy_.reset(active_entry_proxy.release());
}
int SimpleEntryImpl::OpenEntry(Entry** out_entry,
const CompletionCallback& callback) {
DCHECK(backend_.get());
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_OPEN_CALL);
bool have_index = backend_->index()->initialized();
// This enumeration is used in histograms, add entries only at end.
enum OpenEntryIndexEnum {
INDEX_NOEXIST = 0,
INDEX_MISS = 1,
INDEX_HIT = 2,
INDEX_MAX = 3,
};
OpenEntryIndexEnum open_entry_index_enum = INDEX_NOEXIST;
if (have_index) {
if (backend_->index()->Has(entry_hash_))
open_entry_index_enum = INDEX_HIT;
else
open_entry_index_enum = INDEX_MISS;
}
SIMPLE_CACHE_UMA(ENUMERATION,
"OpenEntryIndexState", cache_type_,
open_entry_index_enum, INDEX_MAX);
// If entry is not known to the index, initiate fast failover to the network.
if (open_entry_index_enum == INDEX_MISS) {
net_log_.AddEventWithNetErrorCode(
net::NetLogEventType::SIMPLE_CACHE_ENTRY_OPEN_END, net::ERR_FAILED);
return net::ERR_FAILED;
}
pending_operations_.push(SimpleEntryOperation::OpenOperation(
this, have_index, callback, out_entry));
RunNextOperationIfNeeded();
return net::ERR_IO_PENDING;
}
int SimpleEntryImpl::CreateEntry(Entry** out_entry,
const CompletionCallback& callback) {
DCHECK(backend_.get());
DCHECK_EQ(entry_hash_, simple_util::GetEntryHashKey(key_));
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_CREATE_CALL);
bool have_index = backend_->index()->initialized();
int ret_value = net::ERR_FAILED;
if (use_optimistic_operations_ &&
state_ == STATE_UNINITIALIZED && pending_operations_.size() == 0) {
net_log_.AddEvent(
net::NetLogEventType::SIMPLE_CACHE_ENTRY_CREATE_OPTIMISTIC);
ReturnEntryToCaller(out_entry);
pending_operations_.push(SimpleEntryOperation::CreateOperation(
this, have_index, CompletionCallback(), static_cast<Entry**>(NULL)));
ret_value = net::OK;
} else {
pending_operations_.push(SimpleEntryOperation::CreateOperation(
this, have_index, callback, out_entry));
ret_value = net::ERR_IO_PENDING;
}
// We insert the entry in the index before creating the entry files in the
// SimpleSynchronousEntry, because this way the worst scenario is when we
// have the entry in the index but we don't have the created files yet, this
// way we never leak files. CreationOperationComplete will remove the entry
// from the index if the creation fails.
backend_->index()->Insert(entry_hash_);
RunNextOperationIfNeeded();
return ret_value;
}
int SimpleEntryImpl::DoomEntry(const CompletionCallback& callback) {
if (doomed_)
return net::OK;
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_DOOM_CALL);
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_DOOM_BEGIN);
MarkAsDoomed();
if (backend_.get())
backend_->OnDoomStart(entry_hash_);
pending_operations_.push(SimpleEntryOperation::DoomOperation(this, callback));
RunNextOperationIfNeeded();
return net::ERR_IO_PENDING;
}
void SimpleEntryImpl::SetKey(const std::string& key) {
key_ = key;
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_SET_KEY,
net::NetLog::StringCallback("key", &key));
}
void SimpleEntryImpl::Doom() {
DoomEntry(CompletionCallback());
}
void SimpleEntryImpl::Close() {
DCHECK(io_thread_checker_.CalledOnValidThread());
DCHECK_LT(0, open_count_);
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_CLOSE_CALL);
if (--open_count_ > 0) {
DCHECK(!HasOneRef());
Release(); // Balanced in ReturnEntryToCaller().
return;
}
pending_operations_.push(SimpleEntryOperation::CloseOperation(this));
DCHECK(!HasOneRef());
Release(); // Balanced in ReturnEntryToCaller().
RunNextOperationIfNeeded();
}
std::string SimpleEntryImpl::GetKey() const {
DCHECK(io_thread_checker_.CalledOnValidThread());
return key_;
}
Time SimpleEntryImpl::GetLastUsed() const {
DCHECK(io_thread_checker_.CalledOnValidThread());
return last_used_;
}
Time SimpleEntryImpl::GetLastModified() const {
DCHECK(io_thread_checker_.CalledOnValidThread());
return last_modified_;
}
int32_t SimpleEntryImpl::GetDataSize(int stream_index) const {
DCHECK(io_thread_checker_.CalledOnValidThread());
DCHECK_LE(0, data_size_[stream_index]);
return data_size_[stream_index];
}
int SimpleEntryImpl::ReadData(int stream_index,
int offset,
net::IOBuffer* buf,
int buf_len,
const CompletionCallback& callback) {
DCHECK(io_thread_checker_.CalledOnValidThread());
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_READ_CALL,
CreateNetLogReadWriteDataCallback(stream_index, offset,
buf_len, false));
}
if (stream_index < 0 || stream_index >= kSimpleEntryStreamCount ||
buf_len < 0) {
if (net_log_.IsCapturing()) {
net_log_.AddEvent(
net::NetLogEventType::SIMPLE_CACHE_ENTRY_READ_END,
CreateNetLogReadWriteCompleteCallback(net::ERR_INVALID_ARGUMENT));
}
RecordReadResult(cache_type_, READ_RESULT_INVALID_ARGUMENT);
return net::ERR_INVALID_ARGUMENT;
}
if (pending_operations_.empty() && (offset >= GetDataSize(stream_index) ||
offset < 0 || !buf_len)) {
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_READ_END,
CreateNetLogReadWriteCompleteCallback(0));
}
RecordReadResult(cache_type_, READ_RESULT_NONBLOCK_EMPTY_RETURN);
return 0;
}
// TODO(clamy): return immediatly when reading from stream 0.
// TODO(felipeg): Optimization: Add support for truly parallel read
// operations.
bool alone_in_queue =
pending_operations_.size() == 0 && state_ == STATE_READY;
pending_operations_.push(SimpleEntryOperation::ReadOperation(
this, stream_index, offset, buf_len, buf, callback, alone_in_queue));
RunNextOperationIfNeeded();
return net::ERR_IO_PENDING;
}
int SimpleEntryImpl::WriteData(int stream_index,
int offset,
net::IOBuffer* buf,
int buf_len,
const CompletionCallback& callback,
bool truncate) {
DCHECK(io_thread_checker_.CalledOnValidThread());
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_WRITE_CALL,
CreateNetLogReadWriteDataCallback(stream_index, offset,
buf_len, truncate));
}
if (stream_index < 0 || stream_index >= kSimpleEntryStreamCount ||
offset < 0 || buf_len < 0) {
if (net_log_.IsCapturing()) {
net_log_.AddEvent(
net::NetLogEventType::SIMPLE_CACHE_ENTRY_WRITE_END,
CreateNetLogReadWriteCompleteCallback(net::ERR_INVALID_ARGUMENT));
}
RecordWriteResult(cache_type_, WRITE_RESULT_INVALID_ARGUMENT);
return net::ERR_INVALID_ARGUMENT;
}
if (backend_.get() && offset + buf_len > backend_->GetMaxFileSize()) {
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_WRITE_END,
CreateNetLogReadWriteCompleteCallback(net::ERR_FAILED));
}
RecordWriteResult(cache_type_, WRITE_RESULT_OVER_MAX_SIZE);
return net::ERR_FAILED;
}
ScopedOperationRunner operation_runner(this);
// Stream 0 data is kept in memory, so can be written immediatly if there are
// no IO operations pending.
if (stream_index == 0 && state_ == STATE_READY &&
pending_operations_.size() == 0)
return SetStream0Data(buf, offset, buf_len, truncate);
// We can only do optimistic Write if there is no pending operations, so
// that we are sure that the next call to RunNextOperationIfNeeded will
// actually run the write operation that sets the stream size. It also
// prevents from previous possibly-conflicting writes that could be stacked
// in the |pending_operations_|. We could optimize this for when we have
// only read operations enqueued.
const bool optimistic =
(use_optimistic_operations_ && state_ == STATE_READY &&
pending_operations_.size() == 0);
CompletionCallback op_callback;
scoped_refptr<net::IOBuffer> op_buf;
int ret_value = net::ERR_FAILED;
if (!optimistic) {
op_buf = buf;
op_callback = callback;
ret_value = net::ERR_IO_PENDING;
} else {
// TODO(gavinp,pasko): For performance, don't use a copy of an IOBuffer
// here to avoid paying the price of the RefCountedThreadSafe atomic
// operations.
if (buf) {
op_buf = new IOBuffer(buf_len);
memcpy(op_buf->data(), buf->data(), buf_len);
}
op_callback = CompletionCallback();
ret_value = buf_len;
if (net_log_.IsCapturing()) {
net_log_.AddEvent(
net::NetLogEventType::SIMPLE_CACHE_ENTRY_WRITE_OPTIMISTIC,
CreateNetLogReadWriteCompleteCallback(buf_len));
}
}
pending_operations_.push(SimpleEntryOperation::WriteOperation(this,
stream_index,
offset,
buf_len,
op_buf.get(),
truncate,
optimistic,
op_callback));
return ret_value;
}
int SimpleEntryImpl::ReadSparseData(int64_t offset,
net::IOBuffer* buf,
int buf_len,
const CompletionCallback& callback) {
DCHECK(io_thread_checker_.CalledOnValidThread());
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_READ_SPARSE_CALL,
CreateNetLogSparseOperationCallback(offset, buf_len));
}
ScopedOperationRunner operation_runner(this);
pending_operations_.push(SimpleEntryOperation::ReadSparseOperation(
this, offset, buf_len, buf, callback));
return net::ERR_IO_PENDING;
}
int SimpleEntryImpl::WriteSparseData(int64_t offset,
net::IOBuffer* buf,
int buf_len,
const CompletionCallback& callback) {
DCHECK(io_thread_checker_.CalledOnValidThread());
if (net_log_.IsCapturing()) {
net_log_.AddEvent(
net::NetLogEventType::SIMPLE_CACHE_ENTRY_WRITE_SPARSE_CALL,
CreateNetLogSparseOperationCallback(offset, buf_len));
}
ScopedOperationRunner operation_runner(this);
pending_operations_.push(SimpleEntryOperation::WriteSparseOperation(
this, offset, buf_len, buf, callback));
return net::ERR_IO_PENDING;
}
int SimpleEntryImpl::GetAvailableRange(int64_t offset,
int len,
int64_t* start,
const CompletionCallback& callback) {
DCHECK(io_thread_checker_.CalledOnValidThread());
ScopedOperationRunner operation_runner(this);
pending_operations_.push(SimpleEntryOperation::GetAvailableRangeOperation(
this, offset, len, start, callback));
return net::ERR_IO_PENDING;
}
bool SimpleEntryImpl::CouldBeSparse() const {
DCHECK(io_thread_checker_.CalledOnValidThread());
// TODO(juliatuttle): Actually check.
return true;
}
void SimpleEntryImpl::CancelSparseIO() {
DCHECK(io_thread_checker_.CalledOnValidThread());
// The Simple Cache does not return distinct objects for the same non-doomed
// entry, so there's no need to coordinate which object is performing sparse
// I/O. Therefore, CancelSparseIO and ReadyForSparseIO succeed instantly.
}
int SimpleEntryImpl::ReadyForSparseIO(const CompletionCallback& callback) {
DCHECK(io_thread_checker_.CalledOnValidThread());
// The simple Cache does not return distinct objects for the same non-doomed
// entry, so there's no need to coordinate which object is performing sparse
// I/O. Therefore, CancelSparseIO and ReadyForSparseIO succeed instantly.
return net::OK;
}
SimpleEntryImpl::~SimpleEntryImpl() {
DCHECK(io_thread_checker_.CalledOnValidThread());
DCHECK_EQ(0U, pending_operations_.size());
DCHECK(state_ == STATE_UNINITIALIZED || state_ == STATE_FAILURE);
DCHECK(!synchronous_entry_);
net_log_.EndEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY);
}
void SimpleEntryImpl::PostClientCallback(const CompletionCallback& callback,
int result) {
if (callback.is_null())
return;
// Note that the callback is posted rather than directly invoked to avoid
// reentrancy issues.
base::ThreadTaskRunnerHandle::Get()->PostTask(
FROM_HERE,
base::Bind(&InvokeCallbackIfBackendIsAlive, backend_, callback, result));
}
void SimpleEntryImpl::MakeUninitialized() {
state_ = STATE_UNINITIALIZED;
std::memset(crc32s_end_offset_, 0, sizeof(crc32s_end_offset_));
std::memset(crc32s_, 0, sizeof(crc32s_));
std::memset(have_written_, 0, sizeof(have_written_));
std::memset(data_size_, 0, sizeof(data_size_));
for (size_t i = 0; i < arraysize(crc_check_state_); ++i) {
crc_check_state_[i] = CRC_CHECK_NEVER_READ_AT_ALL;
}
}
void SimpleEntryImpl::ReturnEntryToCaller(Entry** out_entry) {
DCHECK(out_entry);
++open_count_;
AddRef(); // Balanced in Close()
if (!backend_.get()) {
// This method can be called when an asynchronous operation completed.
// If the backend no longer exists, the callback won't be invoked, and so we
// must close ourselves to avoid leaking. As well, there's no guarantee the
// client-provided pointer (|out_entry|) hasn't been freed, and no point
// dereferencing it, either.
Close();
return;
}
*out_entry = this;
}
void SimpleEntryImpl::MarkAsDoomed() {
doomed_ = true;
if (!backend_.get())
return;
backend_->index()->Remove(entry_hash_);
active_entry_proxy_.reset();
}
void SimpleEntryImpl::RunNextOperationIfNeeded() {
DCHECK(io_thread_checker_.CalledOnValidThread());
SIMPLE_CACHE_UMA(CUSTOM_COUNTS,
"EntryOperationsPending", cache_type_,
pending_operations_.size(), 0, 100, 20);
if (!pending_operations_.empty() && state_ != STATE_IO_PENDING) {
std::unique_ptr<SimpleEntryOperation> operation(
new SimpleEntryOperation(pending_operations_.front()));
pending_operations_.pop();
switch (operation->type()) {
case SimpleEntryOperation::TYPE_OPEN:
OpenEntryInternal(operation->have_index(),
operation->callback(),
operation->out_entry());
break;
case SimpleEntryOperation::TYPE_CREATE:
CreateEntryInternal(operation->have_index(),
operation->callback(),
operation->out_entry());
break;
case SimpleEntryOperation::TYPE_CLOSE:
CloseInternal();
break;
case SimpleEntryOperation::TYPE_READ:
RecordReadIsParallelizable(*operation);
ReadDataInternal(operation->index(),
operation->offset(),
operation->buf(),
operation->length(),
operation->callback());
break;
case SimpleEntryOperation::TYPE_WRITE:
RecordWriteDependencyType(*operation);
WriteDataInternal(operation->index(),
operation->offset(),
operation->buf(),
operation->length(),
operation->callback(),
operation->truncate());
break;
case SimpleEntryOperation::TYPE_READ_SPARSE:
ReadSparseDataInternal(operation->sparse_offset(),
operation->buf(),
operation->length(),
operation->callback());
break;
case SimpleEntryOperation::TYPE_WRITE_SPARSE:
WriteSparseDataInternal(operation->sparse_offset(),
operation->buf(),
operation->length(),
operation->callback());
break;
case SimpleEntryOperation::TYPE_GET_AVAILABLE_RANGE:
GetAvailableRangeInternal(operation->sparse_offset(),
operation->length(),
operation->out_start(),
operation->callback());
break;
case SimpleEntryOperation::TYPE_DOOM:
DoomEntryInternal(operation->callback());
break;
default:
NOTREACHED();
}
// The operation is kept for histograms. Makes sure it does not leak
// resources.
executing_operation_.swap(operation);
executing_operation_->ReleaseReferences();
// |this| may have been deleted.
}
}
void SimpleEntryImpl::OpenEntryInternal(bool have_index,
const CompletionCallback& callback,
Entry** out_entry) {
ScopedOperationRunner operation_runner(this);
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_OPEN_BEGIN);
if (state_ == STATE_READY) {
ReturnEntryToCaller(out_entry);
PostClientCallback(callback, net::OK);
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_OPEN_END,
CreateNetLogSimpleEntryCreationCallback(this, net::OK));
return;
}
if (state_ == STATE_FAILURE) {
PostClientCallback(callback, net::ERR_FAILED);
net_log_.AddEvent(
net::NetLogEventType::SIMPLE_CACHE_ENTRY_OPEN_END,
CreateNetLogSimpleEntryCreationCallback(this, net::ERR_FAILED));
return;
}
DCHECK_EQ(STATE_UNINITIALIZED, state_);
DCHECK(!synchronous_entry_);
state_ = STATE_IO_PENDING;
const base::TimeTicks start_time = base::TimeTicks::Now();
std::unique_ptr<SimpleEntryCreationResults> results(
new SimpleEntryCreationResults(SimpleEntryStat(
last_used_, last_modified_, data_size_, sparse_data_size_)));
Closure task =
base::Bind(&SimpleSynchronousEntry::OpenEntry, cache_type_, path_, key_,
entry_hash_, have_index, results.get());
Closure reply =
base::Bind(&SimpleEntryImpl::CreationOperationComplete, this, callback,
start_time, base::Passed(&results), out_entry,
net::NetLogEventType::SIMPLE_CACHE_ENTRY_OPEN_END);
worker_pool_->PostTaskAndReply(FROM_HERE, task, reply);
}
void SimpleEntryImpl::CreateEntryInternal(bool have_index,
const CompletionCallback& callback,
Entry** out_entry) {
ScopedOperationRunner operation_runner(this);
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_CREATE_BEGIN);
if (state_ != STATE_UNINITIALIZED) {
// There is already an active normal entry.
net_log_.AddEvent(
net::NetLogEventType::SIMPLE_CACHE_ENTRY_CREATE_END,
CreateNetLogSimpleEntryCreationCallback(this, net::ERR_FAILED));
PostClientCallback(callback, net::ERR_FAILED);
return;
}
DCHECK_EQ(STATE_UNINITIALIZED, state_);
DCHECK(!synchronous_entry_);
state_ = STATE_IO_PENDING;
// Since we don't know the correct values for |last_used_| and
// |last_modified_| yet, we make this approximation.
last_used_ = last_modified_ = base::Time::Now();
// If creation succeeds, we should mark all streams to be saved on close.
for (int i = 0; i < kSimpleEntryStreamCount; ++i)
have_written_[i] = true;
const base::TimeTicks start_time = base::TimeTicks::Now();
std::unique_ptr<SimpleEntryCreationResults> results(
new SimpleEntryCreationResults(SimpleEntryStat(
last_used_, last_modified_, data_size_, sparse_data_size_)));
Closure task = base::Bind(&SimpleSynchronousEntry::CreateEntry,
cache_type_,
path_,
key_,
entry_hash_,
have_index,
results.get());
Closure reply =
base::Bind(&SimpleEntryImpl::CreationOperationComplete, this, callback,
start_time, base::Passed(&results), out_entry,
net::NetLogEventType::SIMPLE_CACHE_ENTRY_CREATE_END);
worker_pool_->PostTaskAndReply(FROM_HERE, task, reply);
}
void SimpleEntryImpl::CloseInternal() {
DCHECK(io_thread_checker_.CalledOnValidThread());
typedef SimpleSynchronousEntry::CRCRecord CRCRecord;
std::unique_ptr<std::vector<CRCRecord>> crc32s_to_write(
new std::vector<CRCRecord>());
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_CLOSE_BEGIN);
if (state_ == STATE_READY) {
DCHECK(synchronous_entry_);
state_ = STATE_IO_PENDING;
for (int i = 0; i < kSimpleEntryStreamCount; ++i) {
if (have_written_[i]) {
if (GetDataSize(i) == crc32s_end_offset_[i]) {
int32_t crc = GetDataSize(i) == 0 ? crc32(0, Z_NULL, 0) : crc32s_[i];
crc32s_to_write->push_back(CRCRecord(i, true, crc));
} else {
crc32s_to_write->push_back(CRCRecord(i, false, 0));
}
}
}
} else {
DCHECK(STATE_UNINITIALIZED == state_ || STATE_FAILURE == state_);
}
if (synchronous_entry_) {
Closure task = base::Bind(
&SimpleSynchronousEntry::Close, base::Unretained(synchronous_entry_),
SimpleEntryStat(last_used_, last_modified_, data_size_,
sparse_data_size_),
base::Passed(&crc32s_to_write), base::RetainedRef(stream_0_data_));
Closure reply = base::Bind(&SimpleEntryImpl::CloseOperationComplete, this);
synchronous_entry_ = NULL;
worker_pool_->PostTaskAndReply(FROM_HERE, task, reply);
for (int i = 0; i < kSimpleEntryStreamCount; ++i) {
if (!have_written_[i]) {
SIMPLE_CACHE_UMA(ENUMERATION,
"CheckCRCResult", cache_type_,
crc_check_state_[i], CRC_CHECK_MAX);
}
}
} else {
CloseOperationComplete();
}
}
void SimpleEntryImpl::ReadDataInternal(int stream_index,
int offset,
net::IOBuffer* buf,
int buf_len,
const CompletionCallback& callback) {
DCHECK(io_thread_checker_.CalledOnValidThread());
ScopedOperationRunner operation_runner(this);
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_READ_BEGIN,
CreateNetLogReadWriteDataCallback(stream_index, offset,
buf_len, false));
}
if (state_ == STATE_FAILURE || state_ == STATE_UNINITIALIZED) {
if (!callback.is_null()) {
RecordReadResult(cache_type_, READ_RESULT_BAD_STATE);
// Note that the API states that client-provided callbacks for entry-level
// (i.e. non-backend) operations (e.g. read, write) are invoked even if
// the backend was already destroyed.
base::ThreadTaskRunnerHandle::Get()->PostTask(
FROM_HERE, base::Bind(callback, net::ERR_FAILED));
}
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_READ_END,
CreateNetLogReadWriteCompleteCallback(net::ERR_FAILED));
}
return;
}
DCHECK_EQ(STATE_READY, state_);
if (offset >= GetDataSize(stream_index) || offset < 0 || !buf_len) {
RecordReadResult(cache_type_, READ_RESULT_FAST_EMPTY_RETURN);
// If there is nothing to read, we bail out before setting state_ to
// STATE_IO_PENDING.
if (!callback.is_null())
base::ThreadTaskRunnerHandle::Get()->PostTask(FROM_HERE,
base::Bind(callback, 0));
return;
}
buf_len = std::min(buf_len, GetDataSize(stream_index) - offset);
// Since stream 0 data is kept in memory, it is read immediately.
if (stream_index == 0) {
int ret_value = ReadStream0Data(buf, offset, buf_len);
if (!callback.is_null()) {
base::ThreadTaskRunnerHandle::Get()->PostTask(
FROM_HERE, base::Bind(callback, ret_value));
}
return;
}
state_ = STATE_IO_PENDING;
if (!doomed_ && backend_.get())
backend_->index()->UseIfExists(entry_hash_);
std::unique_ptr<uint32_t> read_crc32(new uint32_t());
std::unique_ptr<int> result(new int());
std::unique_ptr<SimpleEntryStat> entry_stat(new SimpleEntryStat(
last_used_, last_modified_, data_size_, sparse_data_size_));
Closure task = base::Bind(
&SimpleSynchronousEntry::ReadData, base::Unretained(synchronous_entry_),
SimpleSynchronousEntry::EntryOperationData(stream_index, offset, buf_len),
base::RetainedRef(buf), read_crc32.get(), entry_stat.get(), result.get());
Closure reply = base::Bind(&SimpleEntryImpl::ReadOperationComplete,
this,
stream_index,
offset,
callback,
base::Passed(&read_crc32),
base::Passed(&entry_stat),
base::Passed(&result));
worker_pool_->PostTaskAndReply(FROM_HERE, task, reply);
}
void SimpleEntryImpl::WriteDataInternal(int stream_index,
int offset,
net::IOBuffer* buf,
int buf_len,
const CompletionCallback& callback,
bool truncate) {
DCHECK(io_thread_checker_.CalledOnValidThread());
ScopedOperationRunner operation_runner(this);
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_WRITE_BEGIN,
CreateNetLogReadWriteDataCallback(stream_index, offset,
buf_len, truncate));
}
if (state_ == STATE_FAILURE || state_ == STATE_UNINITIALIZED) {
RecordWriteResult(cache_type_, WRITE_RESULT_BAD_STATE);
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_WRITE_END,
CreateNetLogReadWriteCompleteCallback(net::ERR_FAILED));
}
if (!callback.is_null()) {
base::ThreadTaskRunnerHandle::Get()->PostTask(
FROM_HERE, base::Bind(callback, net::ERR_FAILED));
}
// |this| may be destroyed after return here.
return;
}
DCHECK_EQ(STATE_READY, state_);
// Since stream 0 data is kept in memory, it will be written immediatly.
if (stream_index == 0) {
int ret_value = SetStream0Data(buf, offset, buf_len, truncate);
if (!callback.is_null()) {
base::ThreadTaskRunnerHandle::Get()->PostTask(
FROM_HERE, base::Bind(callback, ret_value));
}
return;
}
// Ignore zero-length writes that do not change the file size.
if (buf_len == 0) {
int32_t data_size = data_size_[stream_index];
if (truncate ? (offset == data_size) : (offset <= data_size)) {
RecordWriteResult(cache_type_, WRITE_RESULT_FAST_EMPTY_RETURN);
if (!callback.is_null()) {
base::ThreadTaskRunnerHandle::Get()->PostTask(FROM_HERE,
base::Bind(callback, 0));
}
return;
}
}
state_ = STATE_IO_PENDING;
if (!doomed_ && backend_.get())
backend_->index()->UseIfExists(entry_hash_);
AdvanceCrc(buf, offset, buf_len, stream_index);
// |entry_stat| needs to be initialized before modifying |data_size_|.
std::unique_ptr<SimpleEntryStat> entry_stat(new SimpleEntryStat(
last_used_, last_modified_, data_size_, sparse_data_size_));
if (truncate) {
data_size_[stream_index] = offset + buf_len;
} else {
data_size_[stream_index] = std::max(offset + buf_len,
GetDataSize(stream_index));
}
// Since we don't know the correct values for |last_used_| and
// |last_modified_| yet, we make this approximation.
last_used_ = last_modified_ = base::Time::Now();
have_written_[stream_index] = true;
// Writing on stream 1 affects the placement of stream 0 in the file, the EOF
// record will have to be rewritten.
if (stream_index == 1)
have_written_[0] = true;
std::unique_ptr<int> result(new int());
Closure task = base::Bind(
&SimpleSynchronousEntry::WriteData, base::Unretained(synchronous_entry_),
SimpleSynchronousEntry::EntryOperationData(stream_index, offset, buf_len,
truncate, doomed_),
base::RetainedRef(buf), entry_stat.get(), result.get());
Closure reply = base::Bind(&SimpleEntryImpl::WriteOperationComplete,
this,
stream_index,
callback,
base::Passed(&entry_stat),
base::Passed(&result));
worker_pool_->PostTaskAndReply(FROM_HERE, task, reply);
}
void SimpleEntryImpl::ReadSparseDataInternal(
int64_t sparse_offset,
net::IOBuffer* buf,
int buf_len,
const CompletionCallback& callback) {
DCHECK(io_thread_checker_.CalledOnValidThread());
ScopedOperationRunner operation_runner(this);
if (net_log_.IsCapturing()) {
net_log_.AddEvent(
net::NetLogEventType::SIMPLE_CACHE_ENTRY_READ_SPARSE_BEGIN,
CreateNetLogSparseOperationCallback(sparse_offset, buf_len));
}
DCHECK_EQ(STATE_READY, state_);
state_ = STATE_IO_PENDING;
std::unique_ptr<int> result(new int());
std::unique_ptr<base::Time> last_used(new base::Time());
Closure task = base::Bind(
&SimpleSynchronousEntry::ReadSparseData,
base::Unretained(synchronous_entry_),
SimpleSynchronousEntry::EntryOperationData(sparse_offset, buf_len),
base::RetainedRef(buf), last_used.get(), result.get());
Closure reply = base::Bind(&SimpleEntryImpl::ReadSparseOperationComplete,
this,
callback,
base::Passed(&last_used),
base::Passed(&result));
worker_pool_->PostTaskAndReply(FROM_HERE, task, reply);
}
void SimpleEntryImpl::WriteSparseDataInternal(
int64_t sparse_offset,
net::IOBuffer* buf,
int buf_len,
const CompletionCallback& callback) {
DCHECK(io_thread_checker_.CalledOnValidThread());
ScopedOperationRunner operation_runner(this);
if (net_log_.IsCapturing()) {
net_log_.AddEvent(
net::NetLogEventType::SIMPLE_CACHE_ENTRY_WRITE_SPARSE_BEGIN,
CreateNetLogSparseOperationCallback(sparse_offset, buf_len));
}
DCHECK_EQ(STATE_READY, state_);
state_ = STATE_IO_PENDING;
uint64_t max_sparse_data_size = std::numeric_limits<int64_t>::max();
if (backend_.get()) {
uint64_t max_cache_size = backend_->index()->max_size();
max_sparse_data_size = max_cache_size / kMaxSparseDataSizeDivisor;
}
std::unique_ptr<SimpleEntryStat> entry_stat(new SimpleEntryStat(
last_used_, last_modified_, data_size_, sparse_data_size_));
last_used_ = last_modified_ = base::Time::Now();
std::unique_ptr<int> result(new int());
Closure task = base::Bind(
&SimpleSynchronousEntry::WriteSparseData,
base::Unretained(synchronous_entry_),
SimpleSynchronousEntry::EntryOperationData(sparse_offset, buf_len),
base::RetainedRef(buf), max_sparse_data_size, entry_stat.get(),
result.get());
Closure reply = base::Bind(&SimpleEntryImpl::WriteSparseOperationComplete,
this,
callback,
base::Passed(&entry_stat),
base::Passed(&result));
worker_pool_->PostTaskAndReply(FROM_HERE, task, reply);
}
void SimpleEntryImpl::GetAvailableRangeInternal(
int64_t sparse_offset,
int len,
int64_t* out_start,
const CompletionCallback& callback) {
DCHECK(io_thread_checker_.CalledOnValidThread());
ScopedOperationRunner operation_runner(this);
DCHECK_EQ(STATE_READY, state_);
state_ = STATE_IO_PENDING;
std::unique_ptr<int> result(new int());
Closure task = base::Bind(&SimpleSynchronousEntry::GetAvailableRange,
base::Unretained(synchronous_entry_),
SimpleSynchronousEntry::EntryOperationData(
sparse_offset, len),
out_start,
result.get());
Closure reply = base::Bind(
&SimpleEntryImpl::GetAvailableRangeOperationComplete,
this,
callback,
base::Passed(&result));
worker_pool_->PostTaskAndReply(FROM_HERE, task, reply);
}
void SimpleEntryImpl::DoomEntryInternal(const CompletionCallback& callback) {
if (!backend_) {
// If there's no backend, we want to truncate the files rather than delete
// them. Removing files will update the entry directory's mtime, which will
// likely force a full index rebuild on the next startup; this is clearly an
// undesirable cost. Instead, the lesser evil is to set the entry files to
// length zero, leaving the invalid entry in the index. On the next attempt
// to open the entry, it will fail asynchronously (since the magic numbers
// will not be found), and the files will actually be removed.
PostTaskAndReplyWithResult(
worker_pool_.get(), FROM_HERE,
base::Bind(&SimpleSynchronousEntry::TruncateEntryFiles, path_,
entry_hash_),
base::Bind(&SimpleEntryImpl::DoomOperationComplete, this, callback,
// Return to STATE_FAILURE after dooming, since no operation
// can succeed on the truncated entry files.
STATE_FAILURE));
state_ = STATE_IO_PENDING;
return;
}
PostTaskAndReplyWithResult(
worker_pool_.get(),
FROM_HERE,
base::Bind(&SimpleSynchronousEntry::DoomEntry, path_, entry_hash_),
base::Bind(
&SimpleEntryImpl::DoomOperationComplete, this, callback, state_));
state_ = STATE_IO_PENDING;
}
void SimpleEntryImpl::CreationOperationComplete(
const CompletionCallback& completion_callback,
const base::TimeTicks& start_time,
std::unique_ptr<SimpleEntryCreationResults> in_results,
Entry** out_entry,
net::NetLogEventType end_event_type) {
DCHECK(io_thread_checker_.CalledOnValidThread());
DCHECK_EQ(state_, STATE_IO_PENDING);
DCHECK(in_results);
ScopedOperationRunner operation_runner(this);
SIMPLE_CACHE_UMA(BOOLEAN,
"EntryCreationResult", cache_type_,
in_results->result == net::OK);
if (in_results->result != net::OK) {
if (in_results->result != net::ERR_FILE_EXISTS)
MarkAsDoomed();
net_log_.AddEventWithNetErrorCode(end_event_type, net::ERR_FAILED);
PostClientCallback(completion_callback, net::ERR_FAILED);
MakeUninitialized();
return;
}
// If out_entry is NULL, it means we already called ReturnEntryToCaller from
// the optimistic Create case.
if (out_entry)
ReturnEntryToCaller(out_entry);
state_ = STATE_READY;
synchronous_entry_ = in_results->sync_entry;
if (in_results->stream_0_data.get()) {
stream_0_data_ = in_results->stream_0_data;
// The crc was read in SimpleSynchronousEntry.
crc_check_state_[0] = CRC_CHECK_DONE;
crc32s_[0] = in_results->stream_0_crc32;
crc32s_end_offset_[0] = in_results->entry_stat.data_size(0);
}
// If this entry was opened by hash, key_ could still be empty. If so, update
// it with the key read from the synchronous entry.
if (key_.empty()) {
SetKey(synchronous_entry_->key());
} else {
// This should only be triggered when creating an entry. In the open case
// the key is either copied from the arguments to open, or checked
// in the synchronous entry.
DCHECK_EQ(key_, synchronous_entry_->key());
}
UpdateDataFromEntryStat(in_results->entry_stat);
SIMPLE_CACHE_UMA(TIMES,
"EntryCreationTime", cache_type_,
(base::TimeTicks::Now() - start_time));
AdjustOpenEntryCountBy(cache_type_, 1);
net_log_.AddEvent(end_event_type);
PostClientCallback(completion_callback, net::OK);
}
void SimpleEntryImpl::EntryOperationComplete(
const CompletionCallback& completion_callback,
const SimpleEntryStat& entry_stat,
std::unique_ptr<int> result) {
DCHECK(io_thread_checker_.CalledOnValidThread());
DCHECK(synchronous_entry_);
DCHECK_EQ(STATE_IO_PENDING, state_);
DCHECK(result);
if (*result < 0) {
state_ = STATE_FAILURE;
MarkAsDoomed();
} else {
state_ = STATE_READY;
UpdateDataFromEntryStat(entry_stat);
}
if (!completion_callback.is_null()) {
base::ThreadTaskRunnerHandle::Get()->PostTask(
FROM_HERE, base::Bind(completion_callback, *result));
}
RunNextOperationIfNeeded();
}
void SimpleEntryImpl::ReadOperationComplete(
int stream_index,
int offset,
const CompletionCallback& completion_callback,
std::unique_ptr<uint32_t> read_crc32,
std::unique_ptr<SimpleEntryStat> entry_stat,
std::unique_ptr<int> result) {
DCHECK(io_thread_checker_.CalledOnValidThread());
DCHECK(synchronous_entry_);
DCHECK_EQ(STATE_IO_PENDING, state_);
DCHECK(read_crc32);
DCHECK(result);
if (*result > 0 &&
crc_check_state_[stream_index] == CRC_CHECK_NEVER_READ_AT_ALL) {
crc_check_state_[stream_index] = CRC_CHECK_NEVER_READ_TO_END;
}
if (*result > 0 && crc32s_end_offset_[stream_index] == offset) {
uint32_t current_crc =
offset == 0 ? crc32(0, Z_NULL, 0) : crc32s_[stream_index];
crc32s_[stream_index] = crc32_combine(current_crc, *read_crc32, *result);
crc32s_end_offset_[stream_index] += *result;
if (!have_written_[stream_index] &&
GetDataSize(stream_index) == crc32s_end_offset_[stream_index]) {
// We have just read a file from start to finish, and so we have
// computed a crc of the entire file. We can check it now. If a cache
// entry has a single reader, the normal pattern is to read from start
// to finish.
net_log_.AddEvent(
net::NetLogEventType::SIMPLE_CACHE_ENTRY_CHECKSUM_BEGIN);
std::unique_ptr<int> new_result(new int());
Closure task = base::Bind(&SimpleSynchronousEntry::CheckEOFRecord,
base::Unretained(synchronous_entry_),
stream_index,
*entry_stat,
crc32s_[stream_index],
new_result.get());
Closure reply = base::Bind(&SimpleEntryImpl::ChecksumOperationComplete,
this, *result, stream_index,
completion_callback,
base::Passed(&new_result));
worker_pool_->PostTaskAndReply(FROM_HERE, task, reply);
crc_check_state_[stream_index] = CRC_CHECK_DONE;
return;
}
}
if (*result < 0) {
crc32s_end_offset_[stream_index] = 0;
}
if (*result < 0) {
RecordReadResult(cache_type_, READ_RESULT_SYNC_READ_FAILURE);
} else {
RecordReadResult(cache_type_, READ_RESULT_SUCCESS);
if (crc_check_state_[stream_index] == CRC_CHECK_NEVER_READ_TO_END &&
offset + *result == GetDataSize(stream_index)) {
crc_check_state_[stream_index] = CRC_CHECK_NOT_DONE;
}
}
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_READ_END,
CreateNetLogReadWriteCompleteCallback(*result));
}
EntryOperationComplete(completion_callback, *entry_stat, std::move(result));
}
void SimpleEntryImpl::WriteOperationComplete(
int stream_index,
const CompletionCallback& completion_callback,
std::unique_ptr<SimpleEntryStat> entry_stat,
std::unique_ptr<int> result) {
if (*result >= 0)
RecordWriteResult(cache_type_, WRITE_RESULT_SUCCESS);
else
RecordWriteResult(cache_type_, WRITE_RESULT_SYNC_WRITE_FAILURE);
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_WRITE_END,
CreateNetLogReadWriteCompleteCallback(*result));
}
if (*result < 0) {
crc32s_end_offset_[stream_index] = 0;
}
EntryOperationComplete(completion_callback, *entry_stat, std::move(result));
}
void SimpleEntryImpl::ReadSparseOperationComplete(
const CompletionCallback& completion_callback,
std::unique_ptr<base::Time> last_used,
std::unique_ptr<int> result) {
DCHECK(io_thread_checker_.CalledOnValidThread());
DCHECK(synchronous_entry_);
DCHECK(result);
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_READ_SPARSE_END,
CreateNetLogReadWriteCompleteCallback(*result));
}
SimpleEntryStat entry_stat(*last_used, last_modified_, data_size_,
sparse_data_size_);
EntryOperationComplete(completion_callback, entry_stat, std::move(result));
}
void SimpleEntryImpl::WriteSparseOperationComplete(
const CompletionCallback& completion_callback,
std::unique_ptr<SimpleEntryStat> entry_stat,
std::unique_ptr<int> result) {
DCHECK(io_thread_checker_.CalledOnValidThread());
DCHECK(synchronous_entry_);
DCHECK(result);
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_WRITE_SPARSE_END,
CreateNetLogReadWriteCompleteCallback(*result));
}
EntryOperationComplete(completion_callback, *entry_stat, std::move(result));
}
void SimpleEntryImpl::GetAvailableRangeOperationComplete(
const CompletionCallback& completion_callback,
std::unique_ptr<int> result) {
DCHECK(io_thread_checker_.CalledOnValidThread());
DCHECK(synchronous_entry_);
DCHECK(result);
SimpleEntryStat entry_stat(last_used_, last_modified_, data_size_,
sparse_data_size_);
EntryOperationComplete(completion_callback, entry_stat, std::move(result));
}
void SimpleEntryImpl::DoomOperationComplete(
const CompletionCallback& callback,
State state_to_restore,
int result) {
state_ = state_to_restore;
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_DOOM_END);
PostClientCallback(callback, result);
RunNextOperationIfNeeded();
if (backend_)
backend_->OnDoomComplete(entry_hash_);
}
void SimpleEntryImpl::ChecksumOperationComplete(
int orig_result,
int stream_index,
const CompletionCallback& completion_callback,
std::unique_ptr<int> result) {
DCHECK(io_thread_checker_.CalledOnValidThread());
DCHECK(synchronous_entry_);
DCHECK_EQ(STATE_IO_PENDING, state_);
DCHECK(result);
if (net_log_.IsCapturing()) {
net_log_.AddEventWithNetErrorCode(
net::NetLogEventType::SIMPLE_CACHE_ENTRY_CHECKSUM_END, *result);
}
if (*result == net::OK) {
*result = orig_result;
if (orig_result >= 0)
RecordReadResult(cache_type_, READ_RESULT_SUCCESS);
else
RecordReadResult(cache_type_, READ_RESULT_SYNC_READ_FAILURE);
} else {
RecordReadResult(cache_type_, READ_RESULT_SYNC_CHECKSUM_FAILURE);
}
if (net_log_.IsCapturing()) {
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_READ_END,
CreateNetLogReadWriteCompleteCallback(*result));
}
SimpleEntryStat entry_stat(last_used_, last_modified_, data_size_,
sparse_data_size_);
EntryOperationComplete(completion_callback, entry_stat, std::move(result));
}
void SimpleEntryImpl::CloseOperationComplete() {
DCHECK(!synchronous_entry_);
DCHECK_EQ(0, open_count_);
DCHECK(STATE_IO_PENDING == state_ || STATE_FAILURE == state_ ||
STATE_UNINITIALIZED == state_);
net_log_.AddEvent(net::NetLogEventType::SIMPLE_CACHE_ENTRY_CLOSE_END);
AdjustOpenEntryCountBy(cache_type_, -1);
MakeUninitialized();
RunNextOperationIfNeeded();
}
void SimpleEntryImpl::UpdateDataFromEntryStat(
const SimpleEntryStat& entry_stat) {
DCHECK(io_thread_checker_.CalledOnValidThread());
DCHECK(synchronous_entry_);
DCHECK_EQ(STATE_READY, state_);
last_used_ = entry_stat.last_used();
last_modified_ = entry_stat.last_modified();
for (int i = 0; i < kSimpleEntryStreamCount; ++i) {
data_size_[i] = entry_stat.data_size(i);
}
sparse_data_size_ = entry_stat.sparse_data_size();
if (!doomed_ && backend_.get()) {
backend_->index()->UpdateEntrySize(
entry_hash_, base::checked_cast<uint32_t>(GetDiskUsage()));
}
}
int64_t SimpleEntryImpl::GetDiskUsage() const {
int64_t file_size = 0;
for (int i = 0; i < kSimpleEntryStreamCount; ++i) {
file_size +=
simple_util::GetFileSizeFromDataSize(key_.size(), data_size_[i]);
}
file_size += sparse_data_size_;
return file_size;
}
void SimpleEntryImpl::RecordReadIsParallelizable(
const SimpleEntryOperation& operation) const {
if (!executing_operation_)
return;
// Used in histograms, please only add entries at the end.
enum ReadDependencyType {
// READ_STANDALONE = 0, Deprecated.
READ_FOLLOWS_READ = 1,
READ_FOLLOWS_CONFLICTING_WRITE = 2,
READ_FOLLOWS_NON_CONFLICTING_WRITE = 3,
READ_FOLLOWS_OTHER = 4,
READ_ALONE_IN_QUEUE = 5,
READ_DEPENDENCY_TYPE_MAX = 6,
};
ReadDependencyType type = READ_FOLLOWS_OTHER;
if (operation.alone_in_queue()) {
type = READ_ALONE_IN_QUEUE;
} else if (executing_operation_->type() == SimpleEntryOperation::TYPE_READ) {
type = READ_FOLLOWS_READ;
} else if (executing_operation_->type() == SimpleEntryOperation::TYPE_WRITE) {
if (executing_operation_->ConflictsWith(operation))
type = READ_FOLLOWS_CONFLICTING_WRITE;
else
type = READ_FOLLOWS_NON_CONFLICTING_WRITE;
}
SIMPLE_CACHE_UMA(ENUMERATION,
"ReadIsParallelizable", cache_type_,
type, READ_DEPENDENCY_TYPE_MAX);
}
void SimpleEntryImpl::RecordWriteDependencyType(
const SimpleEntryOperation& operation) const {
if (!executing_operation_)
return;
// Used in histograms, please only add entries at the end.
enum WriteDependencyType {
WRITE_OPTIMISTIC = 0,
WRITE_FOLLOWS_CONFLICTING_OPTIMISTIC = 1,
WRITE_FOLLOWS_NON_CONFLICTING_OPTIMISTIC = 2,
WRITE_FOLLOWS_CONFLICTING_WRITE = 3,
WRITE_FOLLOWS_NON_CONFLICTING_WRITE = 4,
WRITE_FOLLOWS_CONFLICTING_READ = 5,
WRITE_FOLLOWS_NON_CONFLICTING_READ = 6,
WRITE_FOLLOWS_OTHER = 7,
WRITE_DEPENDENCY_TYPE_MAX = 8,
};
WriteDependencyType type = WRITE_FOLLOWS_OTHER;
if (operation.optimistic()) {
type = WRITE_OPTIMISTIC;
} else if (executing_operation_->type() == SimpleEntryOperation::TYPE_READ ||
executing_operation_->type() == SimpleEntryOperation::TYPE_WRITE) {
bool conflicting = executing_operation_->ConflictsWith(operation);
if (executing_operation_->type() == SimpleEntryOperation::TYPE_READ) {
type = conflicting ? WRITE_FOLLOWS_CONFLICTING_READ
: WRITE_FOLLOWS_NON_CONFLICTING_READ;
} else if (executing_operation_->optimistic()) {
type = conflicting ? WRITE_FOLLOWS_CONFLICTING_OPTIMISTIC
: WRITE_FOLLOWS_NON_CONFLICTING_OPTIMISTIC;
} else {
type = conflicting ? WRITE_FOLLOWS_CONFLICTING_WRITE
: WRITE_FOLLOWS_NON_CONFLICTING_WRITE;
}
}
SIMPLE_CACHE_UMA(ENUMERATION,
"WriteDependencyType", cache_type_,
type, WRITE_DEPENDENCY_TYPE_MAX);
}
int SimpleEntryImpl::ReadStream0Data(net::IOBuffer* buf,
int offset,
int buf_len) {
if (buf_len < 0) {
RecordReadResult(cache_type_, READ_RESULT_SYNC_READ_FAILURE);
return 0;
}
memcpy(buf->data(), stream_0_data_->data() + offset, buf_len);
UpdateDataFromEntryStat(
SimpleEntryStat(base::Time::Now(), last_modified_, data_size_,
sparse_data_size_));
RecordReadResult(cache_type_, READ_RESULT_SUCCESS);
return buf_len;
}
int SimpleEntryImpl::SetStream0Data(net::IOBuffer* buf,
int offset,
int buf_len,
bool truncate) {
// Currently, stream 0 is only used for HTTP headers, and always writes them
// with a single, truncating write. Detect these writes and record the size
// changes of the headers. Also, support writes to stream 0 that have
// different access patterns, as required by the API contract.
// All other clients of the Simple Cache are encouraged to use stream 1.
have_written_[0] = true;
int data_size = GetDataSize(0);
if (offset == 0 && truncate) {
RecordHeaderSizeChange(cache_type_, data_size, buf_len);
stream_0_data_->SetCapacity(buf_len);
memcpy(stream_0_data_->data(), buf->data(), buf_len);
data_size_[0] = buf_len;
} else {
RecordUnexpectedStream0Write(cache_type_);
const int buffer_size =
truncate ? offset + buf_len : std::max(offset + buf_len, data_size);
stream_0_data_->SetCapacity(buffer_size);
// If |stream_0_data_| was extended, the extension until offset needs to be
// zero-filled.
const int fill_size = offset <= data_size ? 0 : offset - data_size;
if (fill_size > 0)
memset(stream_0_data_->data() + data_size, 0, fill_size);
if (buf)
memcpy(stream_0_data_->data() + offset, buf->data(), buf_len);
data_size_[0] = buffer_size;
}
base::Time modification_time = base::Time::Now();
AdvanceCrc(buf, offset, buf_len, 0);
UpdateDataFromEntryStat(
SimpleEntryStat(modification_time, modification_time, data_size_,
sparse_data_size_));
RecordWriteResult(cache_type_, WRITE_RESULT_SUCCESS);
return buf_len;
}
void SimpleEntryImpl::AdvanceCrc(net::IOBuffer* buffer,
int offset,
int length,
int stream_index) {
// It is easy to incrementally compute the CRC from [0 .. |offset + buf_len|)
// if |offset == 0| or we have already computed the CRC for [0 .. offset).
// We rely on most write operations being sequential, start to end to compute
// the crc of the data. When we write to an entry and close without having
// done a sequential write, we don't check the CRC on read.
if (offset == 0 || crc32s_end_offset_[stream_index] == offset) {
uint32_t initial_crc =
(offset != 0) ? crc32s_[stream_index] : crc32(0, Z_NULL, 0);
if (length > 0) {
crc32s_[stream_index] = crc32(
initial_crc, reinterpret_cast<const Bytef*>(buffer->data()), length);
}
crc32s_end_offset_[stream_index] = offset + length;
} else if (offset < crc32s_end_offset_[stream_index]) {
// If a range for which the crc32 was already computed is rewritten, the
// computation of the crc32 need to start from 0 again.
crc32s_end_offset_[stream_index] = 0;
}
}
} // namespace disk_cache