mirror of
https://github.com/klzgrad/naiveproxy.git
synced 2024-11-24 14:26:09 +03:00
481 lines
17 KiB
C++
481 lines
17 KiB
C++
// Copyright 2014 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 "base/memory/discardable_shared_memory.h"
|
|
|
|
#include <stdint.h>
|
|
|
|
#include <algorithm>
|
|
|
|
#include "base/atomicops.h"
|
|
#include "base/bits.h"
|
|
#include "base/logging.h"
|
|
#include "base/memory/shared_memory_tracker.h"
|
|
#include "base/numerics/safe_math.h"
|
|
#include "base/process/process_metrics.h"
|
|
#include "base/trace_event/memory_allocator_dump.h"
|
|
#include "base/trace_event/process_memory_dump.h"
|
|
#include "build/build_config.h"
|
|
|
|
#if defined(OS_POSIX) && !defined(OS_NACL)
|
|
// For madvise() which is available on all POSIX compatible systems.
|
|
#include <sys/mman.h>
|
|
#endif
|
|
|
|
#if defined(OS_ANDROID)
|
|
#include "third_party/ashmem/ashmem.h"
|
|
#endif
|
|
|
|
#if defined(OS_WIN)
|
|
#include "base/win/windows_version.h"
|
|
#endif
|
|
|
|
namespace base {
|
|
namespace {
|
|
|
|
// Use a machine-sized pointer as atomic type. It will use the Atomic32 or
|
|
// Atomic64 routines, depending on the architecture.
|
|
typedef intptr_t AtomicType;
|
|
typedef uintptr_t UAtomicType;
|
|
|
|
// Template specialization for timestamp serialization/deserialization. This
|
|
// is used to serialize timestamps using Unix time on systems where AtomicType
|
|
// does not have enough precision to contain a timestamp in the standard
|
|
// serialized format.
|
|
template <int>
|
|
Time TimeFromWireFormat(int64_t value);
|
|
template <int>
|
|
int64_t TimeToWireFormat(Time time);
|
|
|
|
// Serialize to Unix time when using 4-byte wire format.
|
|
// Note: 19 January 2038, this will cease to work.
|
|
template <>
|
|
Time ALLOW_UNUSED_TYPE TimeFromWireFormat<4>(int64_t value) {
|
|
return value ? Time::UnixEpoch() + TimeDelta::FromSeconds(value) : Time();
|
|
}
|
|
template <>
|
|
int64_t ALLOW_UNUSED_TYPE TimeToWireFormat<4>(Time time) {
|
|
return time > Time::UnixEpoch() ? (time - Time::UnixEpoch()).InSeconds() : 0;
|
|
}
|
|
|
|
// Standard serialization format when using 8-byte wire format.
|
|
template <>
|
|
Time ALLOW_UNUSED_TYPE TimeFromWireFormat<8>(int64_t value) {
|
|
return Time::FromInternalValue(value);
|
|
}
|
|
template <>
|
|
int64_t ALLOW_UNUSED_TYPE TimeToWireFormat<8>(Time time) {
|
|
return time.ToInternalValue();
|
|
}
|
|
|
|
struct SharedState {
|
|
enum LockState { UNLOCKED = 0, LOCKED = 1 };
|
|
|
|
explicit SharedState(AtomicType ivalue) { value.i = ivalue; }
|
|
SharedState(LockState lock_state, Time timestamp) {
|
|
int64_t wire_timestamp = TimeToWireFormat<sizeof(AtomicType)>(timestamp);
|
|
DCHECK_GE(wire_timestamp, 0);
|
|
DCHECK_EQ(lock_state & ~1, 0);
|
|
value.u = (static_cast<UAtomicType>(wire_timestamp) << 1) | lock_state;
|
|
}
|
|
|
|
LockState GetLockState() const { return static_cast<LockState>(value.u & 1); }
|
|
|
|
Time GetTimestamp() const {
|
|
return TimeFromWireFormat<sizeof(AtomicType)>(value.u >> 1);
|
|
}
|
|
|
|
// Bit 1: Lock state. Bit is set when locked.
|
|
// Bit 2..sizeof(AtomicType)*8: Usage timestamp. NULL time when locked or
|
|
// purged.
|
|
union {
|
|
AtomicType i;
|
|
UAtomicType u;
|
|
} value;
|
|
};
|
|
|
|
// Shared state is stored at offset 0 in shared memory segments.
|
|
SharedState* SharedStateFromSharedMemory(const SharedMemory& shared_memory) {
|
|
DCHECK(shared_memory.memory());
|
|
return static_cast<SharedState*>(shared_memory.memory());
|
|
}
|
|
|
|
// Round up |size| to a multiple of page size.
|
|
size_t AlignToPageSize(size_t size) {
|
|
return bits::Align(size, base::GetPageSize());
|
|
}
|
|
|
|
// LockPages/UnlockPages are platform-native discardable page management
|
|
// helper functions. Both expect |offset| to be specified relative to the
|
|
// base address at which |memory| is mapped, and that |offset| and |length|
|
|
// are page-aligned by the caller.
|
|
|
|
// Returns SUCCESS on platforms which do not support discardable pages.
|
|
DiscardableSharedMemory::LockResult LockPages(const SharedMemory& memory,
|
|
size_t offset,
|
|
size_t length) {
|
|
#if defined(OS_ANDROID)
|
|
SharedMemoryHandle handle = memory.handle();
|
|
if (handle.IsValid()) {
|
|
int pin_result = ashmem_pin_region(handle.GetHandle(), offset, length);
|
|
if (pin_result == ASHMEM_WAS_PURGED)
|
|
return DiscardableSharedMemory::PURGED;
|
|
if (pin_result < 0)
|
|
return DiscardableSharedMemory::FAILED;
|
|
}
|
|
#endif
|
|
return DiscardableSharedMemory::SUCCESS;
|
|
}
|
|
|
|
// UnlockPages() is a no-op on platforms not supporting discardable pages.
|
|
void UnlockPages(const SharedMemory& memory, size_t offset, size_t length) {
|
|
#if defined(OS_ANDROID)
|
|
SharedMemoryHandle handle = memory.handle();
|
|
if (handle.IsValid()) {
|
|
int unpin_result = ashmem_unpin_region(handle.GetHandle(), offset, length);
|
|
DCHECK_EQ(0, unpin_result);
|
|
}
|
|
#endif
|
|
}
|
|
|
|
} // namespace
|
|
|
|
DiscardableSharedMemory::DiscardableSharedMemory()
|
|
: mapped_size_(0), locked_page_count_(0) {
|
|
}
|
|
|
|
DiscardableSharedMemory::DiscardableSharedMemory(
|
|
SharedMemoryHandle shared_memory_handle)
|
|
: shared_memory_(shared_memory_handle, false),
|
|
mapped_size_(0),
|
|
locked_page_count_(0) {
|
|
}
|
|
|
|
DiscardableSharedMemory::~DiscardableSharedMemory() {
|
|
}
|
|
|
|
bool DiscardableSharedMemory::CreateAndMap(size_t size) {
|
|
CheckedNumeric<size_t> checked_size = size;
|
|
checked_size += AlignToPageSize(sizeof(SharedState));
|
|
if (!checked_size.IsValid())
|
|
return false;
|
|
|
|
if (!shared_memory_.CreateAndMapAnonymous(checked_size.ValueOrDie()))
|
|
return false;
|
|
|
|
mapped_size_ =
|
|
shared_memory_.mapped_size() - AlignToPageSize(sizeof(SharedState));
|
|
|
|
locked_page_count_ = AlignToPageSize(mapped_size_) / base::GetPageSize();
|
|
#if DCHECK_IS_ON()
|
|
for (size_t page = 0; page < locked_page_count_; ++page)
|
|
locked_pages_.insert(page);
|
|
#endif
|
|
|
|
DCHECK(last_known_usage_.is_null());
|
|
SharedState new_state(SharedState::LOCKED, Time());
|
|
subtle::Release_Store(&SharedStateFromSharedMemory(shared_memory_)->value.i,
|
|
new_state.value.i);
|
|
return true;
|
|
}
|
|
|
|
bool DiscardableSharedMemory::Map(size_t size) {
|
|
if (!shared_memory_.Map(AlignToPageSize(sizeof(SharedState)) + size))
|
|
return false;
|
|
|
|
mapped_size_ =
|
|
shared_memory_.mapped_size() - AlignToPageSize(sizeof(SharedState));
|
|
|
|
locked_page_count_ = AlignToPageSize(mapped_size_) / base::GetPageSize();
|
|
#if DCHECK_IS_ON()
|
|
for (size_t page = 0; page < locked_page_count_; ++page)
|
|
locked_pages_.insert(page);
|
|
#endif
|
|
|
|
return true;
|
|
}
|
|
|
|
bool DiscardableSharedMemory::Unmap() {
|
|
if (!shared_memory_.Unmap())
|
|
return false;
|
|
|
|
locked_page_count_ = 0;
|
|
#if DCHECK_IS_ON()
|
|
locked_pages_.clear();
|
|
#endif
|
|
mapped_size_ = 0;
|
|
return true;
|
|
}
|
|
|
|
DiscardableSharedMemory::LockResult DiscardableSharedMemory::Lock(
|
|
size_t offset, size_t length) {
|
|
DCHECK_EQ(AlignToPageSize(offset), offset);
|
|
DCHECK_EQ(AlignToPageSize(length), length);
|
|
|
|
// Calls to this function must be synchronized properly.
|
|
DFAKE_SCOPED_LOCK(thread_collision_warner_);
|
|
|
|
DCHECK(shared_memory_.memory());
|
|
|
|
// We need to successfully acquire the platform independent lock before
|
|
// individual pages can be locked.
|
|
if (!locked_page_count_) {
|
|
// Return false when instance has been purged or not initialized properly
|
|
// by checking if |last_known_usage_| is NULL.
|
|
if (last_known_usage_.is_null())
|
|
return FAILED;
|
|
|
|
SharedState old_state(SharedState::UNLOCKED, last_known_usage_);
|
|
SharedState new_state(SharedState::LOCKED, Time());
|
|
SharedState result(subtle::Acquire_CompareAndSwap(
|
|
&SharedStateFromSharedMemory(shared_memory_)->value.i,
|
|
old_state.value.i,
|
|
new_state.value.i));
|
|
if (result.value.u != old_state.value.u) {
|
|
// Update |last_known_usage_| in case the above CAS failed because of
|
|
// an incorrect timestamp.
|
|
last_known_usage_ = result.GetTimestamp();
|
|
return FAILED;
|
|
}
|
|
}
|
|
|
|
// Zero for length means "everything onward".
|
|
if (!length)
|
|
length = AlignToPageSize(mapped_size_) - offset;
|
|
|
|
size_t start = offset / base::GetPageSize();
|
|
size_t end = start + length / base::GetPageSize();
|
|
DCHECK_LE(start, end);
|
|
DCHECK_LE(end, AlignToPageSize(mapped_size_) / base::GetPageSize());
|
|
|
|
// Add pages to |locked_page_count_|.
|
|
// Note: Locking a page that is already locked is an error.
|
|
locked_page_count_ += end - start;
|
|
#if DCHECK_IS_ON()
|
|
// Detect incorrect usage by keeping track of exactly what pages are locked.
|
|
for (auto page = start; page < end; ++page) {
|
|
auto result = locked_pages_.insert(page);
|
|
DCHECK(result.second);
|
|
}
|
|
DCHECK_EQ(locked_pages_.size(), locked_page_count_);
|
|
#endif
|
|
|
|
// Always behave as if memory was purged when trying to lock a 0 byte segment.
|
|
if (!length)
|
|
return PURGED;
|
|
|
|
// Ensure that the platform won't discard the required pages.
|
|
return LockPages(shared_memory_,
|
|
AlignToPageSize(sizeof(SharedState)) + offset, length);
|
|
}
|
|
|
|
void DiscardableSharedMemory::Unlock(size_t offset, size_t length) {
|
|
DCHECK_EQ(AlignToPageSize(offset), offset);
|
|
DCHECK_EQ(AlignToPageSize(length), length);
|
|
|
|
// Calls to this function must be synchronized properly.
|
|
DFAKE_SCOPED_LOCK(thread_collision_warner_);
|
|
|
|
// Passing zero for |length| means "everything onward". Note that |length| may
|
|
// still be zero after this calculation, e.g. if |mapped_size_| is zero.
|
|
if (!length)
|
|
length = AlignToPageSize(mapped_size_) - offset;
|
|
|
|
DCHECK(shared_memory_.memory());
|
|
|
|
// Allow the pages to be discarded by the platform, if supported.
|
|
UnlockPages(shared_memory_, AlignToPageSize(sizeof(SharedState)) + offset,
|
|
length);
|
|
|
|
size_t start = offset / base::GetPageSize();
|
|
size_t end = start + length / base::GetPageSize();
|
|
DCHECK_LE(start, end);
|
|
DCHECK_LE(end, AlignToPageSize(mapped_size_) / base::GetPageSize());
|
|
|
|
// Remove pages from |locked_page_count_|.
|
|
// Note: Unlocking a page that is not locked is an error.
|
|
DCHECK_GE(locked_page_count_, end - start);
|
|
locked_page_count_ -= end - start;
|
|
#if DCHECK_IS_ON()
|
|
// Detect incorrect usage by keeping track of exactly what pages are locked.
|
|
for (auto page = start; page < end; ++page) {
|
|
auto erased_count = locked_pages_.erase(page);
|
|
DCHECK_EQ(1u, erased_count);
|
|
}
|
|
DCHECK_EQ(locked_pages_.size(), locked_page_count_);
|
|
#endif
|
|
|
|
// Early out and avoid releasing the platform independent lock if some pages
|
|
// are still locked.
|
|
if (locked_page_count_)
|
|
return;
|
|
|
|
Time current_time = Now();
|
|
DCHECK(!current_time.is_null());
|
|
|
|
SharedState old_state(SharedState::LOCKED, Time());
|
|
SharedState new_state(SharedState::UNLOCKED, current_time);
|
|
// Note: timestamp cannot be NULL as that is a unique value used when
|
|
// locked or purged.
|
|
DCHECK(!new_state.GetTimestamp().is_null());
|
|
// Timestamp precision should at least be accurate to the second.
|
|
DCHECK_EQ((new_state.GetTimestamp() - Time::UnixEpoch()).InSeconds(),
|
|
(current_time - Time::UnixEpoch()).InSeconds());
|
|
SharedState result(subtle::Release_CompareAndSwap(
|
|
&SharedStateFromSharedMemory(shared_memory_)->value.i,
|
|
old_state.value.i,
|
|
new_state.value.i));
|
|
|
|
DCHECK_EQ(old_state.value.u, result.value.u);
|
|
|
|
last_known_usage_ = current_time;
|
|
}
|
|
|
|
void* DiscardableSharedMemory::memory() const {
|
|
return reinterpret_cast<uint8_t*>(shared_memory_.memory()) +
|
|
AlignToPageSize(sizeof(SharedState));
|
|
}
|
|
|
|
bool DiscardableSharedMemory::Purge(Time current_time) {
|
|
// Calls to this function must be synchronized properly.
|
|
DFAKE_SCOPED_LOCK(thread_collision_warner_);
|
|
DCHECK(shared_memory_.memory());
|
|
|
|
SharedState old_state(SharedState::UNLOCKED, last_known_usage_);
|
|
SharedState new_state(SharedState::UNLOCKED, Time());
|
|
SharedState result(subtle::Acquire_CompareAndSwap(
|
|
&SharedStateFromSharedMemory(shared_memory_)->value.i,
|
|
old_state.value.i,
|
|
new_state.value.i));
|
|
|
|
// Update |last_known_usage_| to |current_time| if the memory is locked. This
|
|
// allows the caller to determine if purging failed because last known usage
|
|
// was incorrect or memory was locked. In the second case, the caller should
|
|
// most likely wait for some amount of time before attempting to purge the
|
|
// the memory again.
|
|
if (result.value.u != old_state.value.u) {
|
|
last_known_usage_ = result.GetLockState() == SharedState::LOCKED
|
|
? current_time
|
|
: result.GetTimestamp();
|
|
return false;
|
|
}
|
|
|
|
// The next section will release as much resource as can be done
|
|
// from the purging process, until the client process notices the
|
|
// purge and releases its own references.
|
|
// Note: this memory will not be accessed again. The segment will be
|
|
// freed asynchronously at a later time, so just do the best
|
|
// immediately.
|
|
#if defined(OS_POSIX) && !defined(OS_NACL)
|
|
// Linux and Android provide MADV_REMOVE which is preferred as it has a
|
|
// behavior that can be verified in tests. Other POSIX flavors (MacOSX, BSDs),
|
|
// provide MADV_FREE which has the same result but memory is purged lazily.
|
|
#if defined(OS_LINUX) || defined(OS_ANDROID)
|
|
#define MADV_PURGE_ARGUMENT MADV_REMOVE
|
|
#elif defined(OS_MACOSX)
|
|
// MADV_FREE_REUSABLE is similar to MADV_FREE, but also marks the pages with the
|
|
// reusable bit, which allows both Activity Monitor and memory-infra to
|
|
// correctly track the pages.
|
|
#define MADV_PURGE_ARGUMENT MADV_FREE_REUSABLE
|
|
#else
|
|
#define MADV_PURGE_ARGUMENT MADV_FREE
|
|
#endif
|
|
// Advise the kernel to remove resources associated with purged pages.
|
|
// Subsequent accesses of memory pages will succeed, but might result in
|
|
// zero-fill-on-demand pages.
|
|
if (madvise(reinterpret_cast<char*>(shared_memory_.memory()) +
|
|
AlignToPageSize(sizeof(SharedState)),
|
|
AlignToPageSize(mapped_size_), MADV_PURGE_ARGUMENT)) {
|
|
DPLOG(ERROR) << "madvise() failed";
|
|
}
|
|
#elif defined(OS_WIN)
|
|
if (base::win::GetVersion() >= base::win::VERSION_WIN8_1) {
|
|
// Discard the purged pages, which releases the physical storage (resident
|
|
// memory, compressed or swapped), but leaves them reserved & committed.
|
|
// This does not free commit for use by other applications, but allows the
|
|
// system to avoid compressing/swapping these pages to free physical memory.
|
|
static const auto discard_virtual_memory =
|
|
reinterpret_cast<decltype(&::DiscardVirtualMemory)>(GetProcAddress(
|
|
GetModuleHandle(L"kernel32.dll"), "DiscardVirtualMemory"));
|
|
if (discard_virtual_memory) {
|
|
DWORD discard_result = discard_virtual_memory(
|
|
reinterpret_cast<char*>(shared_memory_.memory()) +
|
|
AlignToPageSize(sizeof(SharedState)),
|
|
AlignToPageSize(mapped_size_));
|
|
if (discard_result != ERROR_SUCCESS) {
|
|
DLOG(DCHECK) << "DiscardVirtualMemory() failed in Purge(): "
|
|
<< logging::SystemErrorCodeToString(discard_result);
|
|
}
|
|
}
|
|
}
|
|
#endif
|
|
|
|
last_known_usage_ = Time();
|
|
return true;
|
|
}
|
|
|
|
bool DiscardableSharedMemory::IsMemoryResident() const {
|
|
DCHECK(shared_memory_.memory());
|
|
|
|
SharedState result(subtle::NoBarrier_Load(
|
|
&SharedStateFromSharedMemory(shared_memory_)->value.i));
|
|
|
|
return result.GetLockState() == SharedState::LOCKED ||
|
|
!result.GetTimestamp().is_null();
|
|
}
|
|
|
|
bool DiscardableSharedMemory::IsMemoryLocked() const {
|
|
DCHECK(shared_memory_.memory());
|
|
|
|
SharedState result(subtle::NoBarrier_Load(
|
|
&SharedStateFromSharedMemory(shared_memory_)->value.i));
|
|
|
|
return result.GetLockState() == SharedState::LOCKED;
|
|
}
|
|
|
|
void DiscardableSharedMemory::Close() {
|
|
shared_memory_.Close();
|
|
}
|
|
|
|
void DiscardableSharedMemory::CreateSharedMemoryOwnershipEdge(
|
|
trace_event::MemoryAllocatorDump* local_segment_dump,
|
|
trace_event::ProcessMemoryDump* pmd,
|
|
bool is_owned) const {
|
|
auto* shared_memory_dump =
|
|
SharedMemoryTracker::GetOrCreateSharedMemoryDump(&shared_memory_, pmd);
|
|
// TODO(ssid): Clean this by a new api to inherit size of parent dump once the
|
|
// we send the full PMD and calculate sizes inside chrome, crbug.com/704203.
|
|
size_t resident_size = shared_memory_dump->GetSizeInternal();
|
|
local_segment_dump->AddScalar(trace_event::MemoryAllocatorDump::kNameSize,
|
|
trace_event::MemoryAllocatorDump::kUnitsBytes,
|
|
resident_size);
|
|
|
|
// By creating an edge with a higher |importance| (w.r.t non-owned dumps)
|
|
// the tracing UI will account the effective size of the segment to the
|
|
// client instead of manager.
|
|
// TODO(ssid): Define better constants in MemoryAllocatorDump for importance
|
|
// values, crbug.com/754793.
|
|
const int kImportance = is_owned ? 2 : 0;
|
|
auto shared_memory_guid = shared_memory_.mapped_id();
|
|
local_segment_dump->AddString("id", "hash", shared_memory_guid.ToString());
|
|
|
|
// Owned discardable segments which are allocated by client process, could
|
|
// have been cleared by the discardable manager. So, the segment need not
|
|
// exist in memory and weak dumps are created to indicate the UI that the dump
|
|
// should exist only if the manager also created the global dump edge.
|
|
if (is_owned) {
|
|
pmd->CreateWeakSharedMemoryOwnershipEdge(local_segment_dump->guid(),
|
|
shared_memory_guid, kImportance);
|
|
} else {
|
|
pmd->CreateSharedMemoryOwnershipEdge(local_segment_dump->guid(),
|
|
shared_memory_guid, kImportance);
|
|
}
|
|
}
|
|
|
|
Time DiscardableSharedMemory::Now() const {
|
|
return Time::Now();
|
|
}
|
|
|
|
} // namespace base
|