blob: 070f97057710e669fcf9ca2e96f21b3702a87423 [file] [log] [blame]
// Copyright 2006-2010 the V8 project authors. All rights reserved.
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are
// met:
//
// * Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
// * Redistributions in binary form must reproduce the above
// copyright notice, this list of conditions and the following
// disclaimer in the documentation and/or other materials provided
// with the distribution.
// * Neither the name of Google Inc. nor the names of its
// contributors may be used to endorse or promote products derived
// from this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#ifndef V8_SPACES_INL_H_
#define V8_SPACES_INL_H_
#include "isolate.h"
#include "spaces.h"
#include "v8memory.h"
namespace v8 {
namespace internal {
// -----------------------------------------------------------------------------
// PageIterator
bool PageIterator::has_next() {
return prev_page_ != stop_page_;
}
Page* PageIterator::next() {
ASSERT(has_next());
prev_page_ = (prev_page_ == NULL)
? space_->first_page_
: prev_page_->next_page();
return prev_page_;
}
// -----------------------------------------------------------------------------
// Page
Page* Page::next_page() {
return heap_->isolate()->memory_allocator()->GetNextPage(this);
}
Address Page::AllocationTop() {
PagedSpace* owner = heap_->isolate()->memory_allocator()->PageOwner(this);
return owner->PageAllocationTop(this);
}
Address Page::AllocationWatermark() {
PagedSpace* owner = heap_->isolate()->memory_allocator()->PageOwner(this);
if (this == owner->AllocationTopPage()) {
return owner->top();
}
return address() + AllocationWatermarkOffset();
}
uint32_t Page::AllocationWatermarkOffset() {
return static_cast<uint32_t>((flags_ & kAllocationWatermarkOffsetMask) >>
kAllocationWatermarkOffsetShift);
}
void Page::SetAllocationWatermark(Address allocation_watermark) {
if ((heap_->gc_state() == Heap::SCAVENGE) && IsWatermarkValid()) {
// When iterating intergenerational references during scavenge
// we might decide to promote an encountered young object.
// We will allocate a space for such an object and put it
// into the promotion queue to process it later.
// If space for object was allocated somewhere beyond allocation
// watermark this might cause garbage pointers to appear under allocation
// watermark. To avoid visiting them during dirty regions iteration
// which might be still in progress we store a valid allocation watermark
// value and mark this page as having an invalid watermark.
SetCachedAllocationWatermark(AllocationWatermark());
InvalidateWatermark(true);
}
flags_ = (flags_ & kFlagsMask) |
Offset(allocation_watermark) << kAllocationWatermarkOffsetShift;
ASSERT(AllocationWatermarkOffset()
== static_cast<uint32_t>(Offset(allocation_watermark)));
}
void Page::SetCachedAllocationWatermark(Address allocation_watermark) {
mc_first_forwarded = allocation_watermark;
}
Address Page::CachedAllocationWatermark() {
return mc_first_forwarded;
}
uint32_t Page::GetRegionMarks() {
return dirty_regions_;
}
void Page::SetRegionMarks(uint32_t marks) {
dirty_regions_ = marks;
}
int Page::GetRegionNumberForAddress(Address addr) {
// Each page is divided into 256 byte regions. Each region has a corresponding
// dirty mark bit in the page header. Region can contain intergenerational
// references iff its dirty mark is set.
// A normal 8K page contains exactly 32 regions so all region marks fit
// into 32-bit integer field. To calculate a region number we just divide
// offset inside page by region size.
// A large page can contain more then 32 regions. But we want to avoid
// additional write barrier code for distinguishing between large and normal
// pages so we just ignore the fact that addr points into a large page and
// calculate region number as if addr pointed into a normal 8K page. This way
// we get a region number modulo 32 so for large pages several regions might
// be mapped to a single dirty mark.
ASSERT_PAGE_ALIGNED(this->address());
STATIC_ASSERT((kPageAlignmentMask >> kRegionSizeLog2) < kBitsPerInt);
// We are using masking with kPageAlignmentMask instead of Page::Offset()
// to get an offset to the beginning of 8K page containing addr not to the
// beginning of actual page which can be bigger then 8K.
intptr_t offset_inside_normal_page = OffsetFrom(addr) & kPageAlignmentMask;
return static_cast<int>(offset_inside_normal_page >> kRegionSizeLog2);
}
uint32_t Page::GetRegionMaskForAddress(Address addr) {
return 1 << GetRegionNumberForAddress(addr);
}
uint32_t Page::GetRegionMaskForSpan(Address start, int length_in_bytes) {
uint32_t result = 0;
if (length_in_bytes >= kPageSize) {
result = kAllRegionsDirtyMarks;
} else if (length_in_bytes > 0) {
int start_region = GetRegionNumberForAddress(start);
int end_region =
GetRegionNumberForAddress(start + length_in_bytes - kPointerSize);
uint32_t start_mask = (~0) << start_region;
uint32_t end_mask = ~((~1) << end_region);
result = start_mask & end_mask;
// if end_region < start_region, the mask is ored.
if (result == 0) result = start_mask | end_mask;
}
#ifdef DEBUG
if (FLAG_enable_slow_asserts) {
uint32_t expected = 0;
for (Address a = start; a < start + length_in_bytes; a += kPointerSize) {
expected |= GetRegionMaskForAddress(a);
}
ASSERT(expected == result);
}
#endif
return result;
}
void Page::MarkRegionDirty(Address address) {
SetRegionMarks(GetRegionMarks() | GetRegionMaskForAddress(address));
}
bool Page::IsRegionDirty(Address address) {
return GetRegionMarks() & GetRegionMaskForAddress(address);
}
void Page::ClearRegionMarks(Address start, Address end, bool reaches_limit) {
int rstart = GetRegionNumberForAddress(start);
int rend = GetRegionNumberForAddress(end);
if (reaches_limit) {
end += 1;
}
if ((rend - rstart) == 0) {
return;
}
uint32_t bitmask = 0;
if ((OffsetFrom(start) & kRegionAlignmentMask) == 0
|| (start == ObjectAreaStart())) {
// First region is fully covered
bitmask = 1 << rstart;
}
while (++rstart < rend) {
bitmask |= 1 << rstart;
}
if (bitmask) {
SetRegionMarks(GetRegionMarks() & ~bitmask);
}
}
void Page::FlipMeaningOfInvalidatedWatermarkFlag(Heap* heap) {
heap->page_watermark_invalidated_mark_ ^= 1 << WATERMARK_INVALIDATED;
}
bool Page::IsWatermarkValid() {
return (flags_ & (1 << WATERMARK_INVALIDATED)) !=
heap_->page_watermark_invalidated_mark_;
}
void Page::InvalidateWatermark(bool value) {
if (value) {
flags_ = (flags_ & ~(1 << WATERMARK_INVALIDATED)) |
heap_->page_watermark_invalidated_mark_;
} else {
flags_ =
(flags_ & ~(1 << WATERMARK_INVALIDATED)) |
(heap_->page_watermark_invalidated_mark_ ^
(1 << WATERMARK_INVALIDATED));
}
ASSERT(IsWatermarkValid() == !value);
}
bool Page::GetPageFlag(PageFlag flag) {
return (flags_ & static_cast<intptr_t>(1 << flag)) != 0;
}
void Page::SetPageFlag(PageFlag flag, bool value) {
if (value) {
flags_ |= static_cast<intptr_t>(1 << flag);
} else {
flags_ &= ~static_cast<intptr_t>(1 << flag);
}
}
void Page::ClearPageFlags() {
flags_ = 0;
}
void Page::ClearGCFields() {
InvalidateWatermark(true);
SetAllocationWatermark(ObjectAreaStart());
if (heap_->gc_state() == Heap::SCAVENGE) {
SetCachedAllocationWatermark(ObjectAreaStart());
}
SetRegionMarks(kAllRegionsCleanMarks);
}
bool Page::WasInUseBeforeMC() {
return GetPageFlag(WAS_IN_USE_BEFORE_MC);
}
void Page::SetWasInUseBeforeMC(bool was_in_use) {
SetPageFlag(WAS_IN_USE_BEFORE_MC, was_in_use);
}
bool Page::IsLargeObjectPage() {
return !GetPageFlag(IS_NORMAL_PAGE);
}
void Page::SetIsLargeObjectPage(bool is_large_object_page) {
SetPageFlag(IS_NORMAL_PAGE, !is_large_object_page);
}
bool Page::IsPageExecutable() {
return GetPageFlag(IS_EXECUTABLE);
}
void Page::SetIsPageExecutable(bool is_page_executable) {
SetPageFlag(IS_EXECUTABLE, is_page_executable);
}
// -----------------------------------------------------------------------------
// MemoryAllocator
void MemoryAllocator::ChunkInfo::init(Address a, size_t s, PagedSpace* o) {
address_ = a;
size_ = s;
owner_ = o;
executable_ = (o == NULL) ? NOT_EXECUTABLE : o->executable();
owner_identity_ = (o == NULL) ? FIRST_SPACE : o->identity();
}
bool MemoryAllocator::IsValidChunk(int chunk_id) {
if (!IsValidChunkId(chunk_id)) return false;
ChunkInfo& c = chunks_[chunk_id];
return (c.address() != NULL) && (c.size() != 0) && (c.owner() != NULL);
}
bool MemoryAllocator::IsValidChunkId(int chunk_id) {
return (0 <= chunk_id) && (chunk_id < max_nof_chunks_);
}
bool MemoryAllocator::IsPageInSpace(Page* p, PagedSpace* space) {
ASSERT(p->is_valid());
int chunk_id = GetChunkId(p);
if (!IsValidChunkId(chunk_id)) return false;
ChunkInfo& c = chunks_[chunk_id];
return (c.address() <= p->address()) &&
(p->address() < c.address() + c.size()) &&
(space == c.owner());
}
Page* MemoryAllocator::GetNextPage(Page* p) {
ASSERT(p->is_valid());
intptr_t raw_addr = p->opaque_header & ~Page::kPageAlignmentMask;
return Page::FromAddress(AddressFrom<Address>(raw_addr));
}
int MemoryAllocator::GetChunkId(Page* p) {
ASSERT(p->is_valid());
return static_cast<int>(p->opaque_header & Page::kPageAlignmentMask);
}
void MemoryAllocator::SetNextPage(Page* prev, Page* next) {
ASSERT(prev->is_valid());
int chunk_id = GetChunkId(prev);
ASSERT_PAGE_ALIGNED(next->address());
prev->opaque_header = OffsetFrom(next->address()) | chunk_id;
}
PagedSpace* MemoryAllocator::PageOwner(Page* page) {
int chunk_id = GetChunkId(page);
ASSERT(IsValidChunk(chunk_id));
return chunks_[chunk_id].owner();
}
bool MemoryAllocator::InInitialChunk(Address address) {
if (initial_chunk_ == NULL) return false;
Address start = static_cast<Address>(initial_chunk_->address());
return (start <= address) && (address < start + initial_chunk_->size());
}
#ifdef ENABLE_HEAP_PROTECTION
void MemoryAllocator::Protect(Address start, size_t size) {
OS::Protect(start, size);
}
void MemoryAllocator::Unprotect(Address start,
size_t size,
Executability executable) {
OS::Unprotect(start, size, executable);
}
void MemoryAllocator::ProtectChunkFromPage(Page* page) {
int id = GetChunkId(page);
OS::Protect(chunks_[id].address(), chunks_[id].size());
}
void MemoryAllocator::UnprotectChunkFromPage(Page* page) {
int id = GetChunkId(page);
OS::Unprotect(chunks_[id].address(), chunks_[id].size(),
chunks_[id].owner()->executable() == EXECUTABLE);
}
#endif
// --------------------------------------------------------------------------
// PagedSpace
bool PagedSpace::Contains(Address addr) {
Page* p = Page::FromAddress(addr);
if (!p->is_valid()) return false;
return heap()->isolate()->memory_allocator()->IsPageInSpace(p, this);
}
// Try linear allocation in the page of alloc_info's allocation top. Does
// not contain slow case logic (eg, move to the next page or try free list
// allocation) so it can be used by all the allocation functions and for all
// the paged spaces.
HeapObject* PagedSpace::AllocateLinearly(AllocationInfo* alloc_info,
int size_in_bytes) {
Address current_top = alloc_info->top;
Address new_top = current_top + size_in_bytes;
if (new_top > alloc_info->limit) return NULL;
alloc_info->top = new_top;
ASSERT(alloc_info->VerifyPagedAllocation());
accounting_stats_.AllocateBytes(size_in_bytes);
return HeapObject::FromAddress(current_top);
}
// Raw allocation.
MaybeObject* PagedSpace::AllocateRaw(int size_in_bytes) {
ASSERT(HasBeenSetup());
ASSERT_OBJECT_SIZE(size_in_bytes);
HeapObject* object = AllocateLinearly(&allocation_info_, size_in_bytes);
if (object != NULL) return object;
object = SlowAllocateRaw(size_in_bytes);
if (object != NULL) return object;
return Failure::RetryAfterGC(identity());
}
// Reallocating (and promoting) objects during a compacting collection.
MaybeObject* PagedSpace::MCAllocateRaw(int size_in_bytes) {
ASSERT(HasBeenSetup());
ASSERT_OBJECT_SIZE(size_in_bytes);
HeapObject* object = AllocateLinearly(&mc_forwarding_info_, size_in_bytes);
if (object != NULL) return object;
object = SlowMCAllocateRaw(size_in_bytes);
if (object != NULL) return object;
return Failure::RetryAfterGC(identity());
}
// -----------------------------------------------------------------------------
// LargeObjectChunk
Address LargeObjectChunk::GetStartAddress() {
// Round the chunk address up to the nearest page-aligned address
// and return the heap object in that page.
Page* page = Page::FromAddress(RoundUp(address(), Page::kPageSize));
return page->ObjectAreaStart();
}
void LargeObjectChunk::Free(Executability executable) {
Isolate* isolate =
Page::FromAddress(RoundUp(address(), Page::kPageSize))->heap_->isolate();
isolate->memory_allocator()->FreeRawMemory(address(), size(), executable);
}
// -----------------------------------------------------------------------------
// NewSpace
MaybeObject* NewSpace::AllocateRawInternal(int size_in_bytes,
AllocationInfo* alloc_info) {
Address new_top = alloc_info->top + size_in_bytes;
if (new_top > alloc_info->limit) return Failure::RetryAfterGC();
Object* obj = HeapObject::FromAddress(alloc_info->top);
alloc_info->top = new_top;
#ifdef DEBUG
SemiSpace* space =
(alloc_info == &allocation_info_) ? &to_space_ : &from_space_;
ASSERT(space->low() <= alloc_info->top
&& alloc_info->top <= space->high()
&& alloc_info->limit == space->high());
#endif
return obj;
}
intptr_t LargeObjectSpace::Available() {
return LargeObjectChunk::ObjectSizeFor(
heap()->isolate()->memory_allocator()->Available());
}
template <typename StringType>
void NewSpace::ShrinkStringAtAllocationBoundary(String* string, int length) {
ASSERT(length <= string->length());
ASSERT(string->IsSeqString());
ASSERT(string->address() + StringType::SizeFor(string->length()) ==
allocation_info_.top);
allocation_info_.top =
string->address() + StringType::SizeFor(length);
string->set_length(length);
}
bool FreeListNode::IsFreeListNode(HeapObject* object) {
return object->map() == HEAP->raw_unchecked_byte_array_map()
|| object->map() == HEAP->raw_unchecked_one_pointer_filler_map()
|| object->map() == HEAP->raw_unchecked_two_pointer_filler_map();
}
} } // namespace v8::internal
#endif // V8_SPACES_INL_H_