8200557: OopStorage parallel iteration scales poorly
Summary: Change representation of sequence of all blocks for better scaling.
Reviewed-by: coleenp, eosterlund
--- a/src/hotspot/share/gc/shared/oopStorage.cpp Thu May 03 14:13:20 2018 -0700
+++ b/src/hotspot/share/gc/shared/oopStorage.cpp Thu May 03 17:36:50 2018 -0400
@@ -28,20 +28,22 @@
#include "logging/log.hpp"
#include "logging/logStream.hpp"
#include "memory/allocation.inline.hpp"
-#include "memory/resourceArea.hpp"
#include "runtime/atomic.hpp"
+#include "runtime/globals.hpp"
#include "runtime/handles.inline.hpp"
#include "runtime/mutex.hpp"
#include "runtime/mutexLocker.hpp"
#include "runtime/orderAccess.inline.hpp"
#include "runtime/safepoint.hpp"
#include "runtime/stubRoutines.hpp"
+#include "runtime/thread.hpp"
#include "utilities/align.hpp"
#include "utilities/count_trailing_zeros.hpp"
#include "utilities/debug.hpp"
#include "utilities/globalDefinitions.hpp"
#include "utilities/macros.hpp"
#include "utilities/ostream.hpp"
+#include "utilities/spinYield.hpp"
OopStorage::BlockEntry::BlockEntry() : _prev(NULL), _next(NULL) {}
@@ -108,6 +110,90 @@
}
}
+OopStorage::BlockArray::BlockArray(size_t size) :
+ _size(size),
+ _block_count(0),
+ _refcount(0)
+{}
+
+OopStorage::BlockArray::~BlockArray() {
+ assert(_refcount == 0, "precondition");
+}
+
+OopStorage::BlockArray* OopStorage::BlockArray::create(size_t size, AllocFailType alloc_fail) {
+ size_t size_in_bytes = blocks_offset() + sizeof(Block*) * size;
+ void* mem = NEW_C_HEAP_ARRAY3(char, size_in_bytes, mtGC, CURRENT_PC, alloc_fail);
+ if (mem == NULL) return NULL;
+ return new (mem) BlockArray(size);
+}
+
+void OopStorage::BlockArray::destroy(BlockArray* ba) {
+ ba->~BlockArray();
+ FREE_C_HEAP_ARRAY(char, ba);
+}
+
+size_t OopStorage::BlockArray::size() const {
+ return _size;
+}
+
+size_t OopStorage::BlockArray::block_count() const {
+ return _block_count;
+}
+
+size_t OopStorage::BlockArray::block_count_acquire() const {
+ return OrderAccess::load_acquire(&_block_count);
+}
+
+void OopStorage::BlockArray::increment_refcount() const {
+ int new_value = Atomic::add(1, &_refcount);
+ assert(new_value >= 1, "negative refcount %d", new_value - 1);
+}
+
+bool OopStorage::BlockArray::decrement_refcount() const {
+ int new_value = Atomic::sub(1, &_refcount);
+ assert(new_value >= 0, "negative refcount %d", new_value);
+ return new_value == 0;
+}
+
+bool OopStorage::BlockArray::push(Block* block) {
+ size_t index = _block_count;
+ if (index < _size) {
+ block->set_active_index(index);
+ *block_ptr(index) = block;
+ // Use a release_store to ensure all the setup is complete before
+ // making the block visible.
+ OrderAccess::release_store(&_block_count, index + 1);
+ return true;
+ } else {
+ return false;
+ }
+}
+
+void OopStorage::BlockArray::remove(Block* block) {
+ assert(_block_count > 0, "array is empty");
+ size_t index = block->active_index();
+ assert(*block_ptr(index) == block, "block not present");
+ size_t last_index = _block_count - 1;
+ Block* last_block = *block_ptr(last_index);
+ last_block->set_active_index(index);
+ *block_ptr(index) = last_block;
+ _block_count = last_index;
+}
+
+void OopStorage::BlockArray::copy_from(const BlockArray* from) {
+ assert(_block_count == 0, "array must be empty");
+ size_t count = from->_block_count;
+ assert(count <= _size, "precondition");
+ Block* const* from_ptr = from->block_ptr(0);
+ Block** to_ptr = block_ptr(0);
+ for (size_t i = 0; i < count; ++i) {
+ Block* block = *from_ptr++;
+ assert(block->active_index() == i, "invariant");
+ *to_ptr++ = block;
+ }
+ _block_count = count;
+}
+
// Blocks start with an array of BitsPerWord oop entries. That array
// is divided into conceptual BytesPerWord sections of BitsPerByte
// entries. Blocks are allocated aligned on section boundaries, for
@@ -125,7 +211,7 @@
_allocated_bitmask(0),
_owner(owner),
_memory(memory),
- _active_entry(),
+ _active_index(0),
_allocate_entry(),
_deferred_updates_next(NULL),
_release_refcount(0)
@@ -146,10 +232,6 @@
const_cast<OopStorage* volatile&>(_owner) = NULL;
}
-const OopStorage::BlockEntry& OopStorage::Block::get_active_entry(const Block& block) {
- return block._active_entry;
-}
-
const OopStorage::BlockEntry& OopStorage::Block::get_allocate_entry(const Block& block) {
return block._allocate_entry;
}
@@ -204,6 +286,20 @@
return (base <= ptr) && (ptr < (base + ARRAY_SIZE(_data)));
}
+size_t OopStorage::Block::active_index() const {
+ return _active_index;
+}
+
+void OopStorage::Block::set_active_index(size_t index) {
+ _active_index = index;
+}
+
+size_t OopStorage::Block::active_index_safe(const Block* block) {
+ STATIC_ASSERT(sizeof(intptr_t) == sizeof(block->_active_index));
+ assert(CanUseSafeFetchN(), "precondition");
+ return SafeFetchN((intptr_t*)&block->_active_index, 0);
+}
+
unsigned OopStorage::Block::get_index(const oop* ptr) const {
assert(contains(ptr), PTR_FORMAT " not in block " PTR_FORMAT, p2i(ptr), p2i(this));
return static_cast<unsigned>(ptr - get_pointer(0));
@@ -246,7 +342,7 @@
// This can return a false positive if ptr is not contained by some
// block. For some uses, it is a precondition that ptr is valid,
-// e.g. contained in some block in owner's _active_list. Other uses
+// e.g. contained in some block in owner's _active_array. Other uses
// require additional validation of the result.
OopStorage::Block*
OopStorage::Block::block_for_ptr(const OopStorage* owner, const oop* ptr) {
@@ -280,12 +376,12 @@
// Allocation involves the _allocate_list, which contains a subset of the
// blocks owned by a storage object. This is a doubly-linked list, linked
// through dedicated fields in the blocks. Full blocks are removed from this
-// list, though they are still present in the _active_list. Empty blocks are
+// list, though they are still present in the _active_array. Empty blocks are
// kept at the end of the _allocate_list, to make it easy for empty block
// deletion to find them.
//
// allocate(), and delete_empty_blocks_concurrent() lock the
-// _allocate_mutex while performing any list modifications.
+// _allocate_mutex while performing any list and array modifications.
//
// allocate() and release() update a block's _allocated_bitmask using CAS
// loops. This prevents loss of updates even though release() performs
@@ -299,7 +395,7 @@
//
// release() is performed lock-free. release() first looks up the block for
// the entry, using address alignment to find the enclosing block (thereby
-// avoiding iteration over the _active_list). Once the block has been
+// avoiding iteration over the _active_array). Once the block has been
// determined, its _allocated_bitmask needs to be updated, and its position in
// the _allocate_list may need to be updated. There are two cases:
//
@@ -340,7 +436,7 @@
// Failed to make new block, no other thread made a block
// available while the mutex was released, and didn't get
// one from a deferred update either, so return failure.
- log_info(oopstorage, ref)("%s: failed allocation", name());
+ log_info(oopstorage, ref)("%s: failed block allocation", name());
return NULL;
}
}
@@ -348,17 +444,21 @@
// Add new block to storage.
log_info(oopstorage, blocks)("%s: new block " PTR_FORMAT, name(), p2i(block));
+ // Add new block to the _active_array, growing if needed.
+ if (!_active_array->push(block)) {
+ if (expand_active_array()) {
+ guarantee(_active_array->push(block), "push failed after expansion");
+ } else {
+ log_info(oopstorage, blocks)("%s: failed active array expand", name());
+ Block::delete_block(*block);
+ return NULL;
+ }
+ }
// Add to end of _allocate_list. The mutex release allowed
// other threads to add blocks to the _allocate_list. We prefer
// to allocate from non-empty blocks, to allow empty blocks to
// be deleted.
_allocate_list.push_back(*block);
- // Add to front of _active_list, and then record as the head
- // block, for concurrent iteration protocol.
- _active_list.push_front(*block);
- ++_block_count;
- // Ensure all setup of block is complete before making it visible.
- OrderAccess::release_store(&_active_head, block);
}
block = _allocate_list.head();
}
@@ -383,6 +483,123 @@
return result;
}
+// Create a new, larger, active array with the same content as the
+// current array, and then replace, relinquishing the old array.
+// Return true if the array was successfully expanded, false to
+// indicate allocation failure.
+bool OopStorage::expand_active_array() {
+ assert_lock_strong(_allocate_mutex);
+ BlockArray* old_array = _active_array;
+ size_t new_size = 2 * old_array->size();
+ log_info(oopstorage, blocks)("%s: expand active array " SIZE_FORMAT,
+ name(), new_size);
+ BlockArray* new_array = BlockArray::create(new_size, AllocFailStrategy::RETURN_NULL);
+ if (new_array == NULL) return false;
+ new_array->copy_from(old_array);
+ replace_active_array(new_array);
+ relinquish_block_array(old_array);
+ return true;
+}
+
+OopStorage::ProtectActive::ProtectActive() : _enter(0), _exit() {}
+
+// Begin read-side critical section.
+uint OopStorage::ProtectActive::read_enter() {
+ return Atomic::add(2u, &_enter);
+}
+
+// End read-side critical section.
+void OopStorage::ProtectActive::read_exit(uint enter_value) {
+ Atomic::add(2u, &_exit[enter_value & 1]);
+}
+
+// Wait until all readers that entered the critical section before
+// synchronization have exited that critical section.
+void OopStorage::ProtectActive::write_synchronize() {
+ SpinYield spinner;
+ // Determine old and new exit counters, based on bit0 of the
+ // on-entry _enter counter.
+ uint value = OrderAccess::load_acquire(&_enter);
+ volatile uint* new_ptr = &_exit[(value + 1) & 1];
+ // Atomically change the in-use exit counter to the new counter, by
+ // adding 1 to the _enter counter (flipping bit0 between 0 and 1)
+ // and initializing the new exit counter to that enter value. Note:
+ // The new exit counter is not being used by read operations until
+ // this change succeeds.
+ uint old;
+ do {
+ old = value;
+ *new_ptr = ++value;
+ value = Atomic::cmpxchg(value, &_enter, old);
+ } while (old != value);
+ // Readers that entered the critical section before we changed the
+ // selected exit counter will use the old exit counter. Readers
+ // entering after the change will use the new exit counter. Wait
+ // for all the critical sections started before the change to
+ // complete, e.g. for the value of old_ptr to catch up with old.
+ volatile uint* old_ptr = &_exit[old & 1];
+ while (old != OrderAccess::load_acquire(old_ptr)) {
+ spinner.wait();
+ }
+}
+
+// Make new_array the _active_array. Increments new_array's refcount
+// to account for the new reference. The assignment is atomic wrto
+// obtain_active_array; once this function returns, it is safe for the
+// caller to relinquish the old array.
+void OopStorage::replace_active_array(BlockArray* new_array) {
+ // Caller has the old array that is the current value of _active_array.
+ // Update new_array refcount to account for the new reference.
+ new_array->increment_refcount();
+ // Install new_array, ensuring its initialization is complete first.
+ OrderAccess::release_store(&_active_array, new_array);
+ // Wait for any readers that could read the old array from _active_array.
+ _protect_active.write_synchronize();
+ // All obtain critical sections that could see the old array have
+ // completed, having incremented the refcount of the old array. The
+ // caller can now safely relinquish the old array.
+}
+
+// Atomically (wrto replace_active_array) get the active array and
+// increment its refcount. This provides safe access to the array,
+// even if an allocate operation expands and replaces the value of
+// _active_array. The caller must relinquish the array when done
+// using it.
+OopStorage::BlockArray* OopStorage::obtain_active_array() const {
+ uint enter_value = _protect_active.read_enter();
+ BlockArray* result = OrderAccess::load_acquire(&_active_array);
+ result->increment_refcount();
+ _protect_active.read_exit(enter_value);
+ return result;
+}
+
+// Decrement refcount of array and destroy if refcount is zero.
+void OopStorage::relinquish_block_array(BlockArray* array) const {
+ if (array->decrement_refcount()) {
+ assert(array != _active_array, "invariant");
+ BlockArray::destroy(array);
+ }
+}
+
+class OopStorage::WithActiveArray : public StackObj {
+ const OopStorage* _storage;
+ BlockArray* _active_array;
+
+public:
+ WithActiveArray(const OopStorage* storage) :
+ _storage(storage),
+ _active_array(storage->obtain_active_array())
+ {}
+
+ ~WithActiveArray() {
+ _storage->relinquish_block_array(_active_array);
+ }
+
+ BlockArray& active_array() const {
+ return *_active_array;
+ }
+};
+
OopStorage::Block* OopStorage::find_block_or_null(const oop* ptr) const {
assert(ptr != NULL, "precondition");
return Block::block_for_ptr(this, ptr);
@@ -392,7 +609,6 @@
uintx old_allocated,
const OopStorage* owner,
const void* block) {
- ResourceMark rm;
Log(oopstorage, blocks) log;
LogStream ls(log.debug());
if (is_full_bitmask(old_allocated)) {
@@ -546,20 +762,21 @@
return dup;
}
+const size_t initial_active_array_size = 8;
+
OopStorage::OopStorage(const char* name,
Mutex* allocate_mutex,
Mutex* active_mutex) :
_name(dup_name(name)),
- _active_list(&Block::get_active_entry),
+ _active_array(BlockArray::create(initial_active_array_size)),
_allocate_list(&Block::get_allocate_entry),
- _active_head(NULL),
_deferred_updates(NULL),
_allocate_mutex(allocate_mutex),
_active_mutex(active_mutex),
_allocation_count(0),
- _block_count(0),
_concurrent_iteration_active(false)
{
+ _active_array->increment_refcount();
assert(_active_mutex->rank() < _allocate_mutex->rank(),
"%s: active_mutex must have lower rank than allocate_mutex", _name);
assert(_active_mutex->_safepoint_check_required != Mutex::_safepoint_check_always,
@@ -583,10 +800,13 @@
while ((block = _allocate_list.head()) != NULL) {
_allocate_list.unlink(*block);
}
- while ((block = _active_list.head()) != NULL) {
- _active_list.unlink(*block);
+ bool unreferenced = _active_array->decrement_refcount();
+ assert(unreferenced, "deleting storage while _active_array is referenced");
+ for (size_t i = _active_array->block_count(); 0 < i; ) {
+ block = _active_array->at(--i);
Block::delete_block(*block);
}
+ BlockArray::destroy(_active_array);
FREE_C_HEAP_ARRAY(char, _name);
}
@@ -598,16 +818,13 @@
// Don't interfere with a concurrent iteration.
if (_concurrent_iteration_active) return;
// Delete empty (and otherwise deletable) blocks from end of _allocate_list.
- for (const Block* block = _allocate_list.ctail();
+ for (Block* block = _allocate_list.tail();
(block != NULL) && block->is_deletable();
- block = _allocate_list.ctail()) {
- _active_list.unlink(*block);
+ block = _allocate_list.tail()) {
+ _active_array->remove(block);
_allocate_list.unlink(*block);
delete_empty_block(*block);
- --_block_count;
}
- // Update _active_head, in case current value was in deleted set.
- _active_head = _active_list.head();
}
void OopStorage::delete_empty_blocks_concurrent() {
@@ -616,14 +833,14 @@
// release the mutex across the block deletions. Set an upper bound
// on how many blocks we'll try to release, so other threads can't
// cause an unbounded stay in this function.
- size_t limit = _block_count;
+ size_t limit = block_count();
for (size_t i = 0; i < limit; ++i) {
// Additional updates might become available while we dropped the
// lock. But limit number processed to limit lock duration.
reduce_deferred_updates();
- const Block* block = _allocate_list.ctail();
+ Block* block = _allocate_list.tail();
if ((block == NULL) || !block->is_deletable()) {
// No block to delete, so done. There could be more pending
// deferred updates that could give us more work to do; deal with
@@ -635,12 +852,7 @@
MutexLockerEx aml(_active_mutex, Mutex::_no_safepoint_check_flag);
// Don't interfere with a concurrent iteration.
if (_concurrent_iteration_active) return;
- // Remove block from _active_list, updating head if needed.
- _active_list.unlink(*block);
- --_block_count;
- if (block == _active_head) {
- _active_head = _active_list.head();
- }
+ _active_array->remove(block);
}
// Remove block from _allocate_list and delete it.
_allocate_list.unlink(*block);
@@ -653,18 +865,17 @@
OopStorage::EntryStatus OopStorage::allocation_status(const oop* ptr) const {
const Block* block = find_block_or_null(ptr);
if (block != NULL) {
- // Verify block is a real block. For now, simple linear search.
- // Do something more clever if this is a performance bottleneck.
+ // Prevent block deletion and _active_array modification.
MutexLockerEx ml(_allocate_mutex, Mutex::_no_safepoint_check_flag);
- for (const Block* check_block = _active_list.chead();
- check_block != NULL;
- check_block = _active_list.next(*check_block)) {
- if (check_block == block) {
- if ((block->allocated_bitmask() & block->bitmask_for_entry(ptr)) != 0) {
- return ALLOCATED_ENTRY;
- } else {
- return UNALLOCATED_ENTRY;
- }
+ // Block could be a false positive, so get index carefully.
+ size_t index = Block::active_index_safe(block);
+ if ((index < _active_array->block_count()) &&
+ (block == _active_array->at(index)) &&
+ block->contains(ptr)) {
+ if ((block->allocated_bitmask() & block->bitmask_for_entry(ptr)) != 0) {
+ return ALLOCATED_ENTRY;
+ } else {
+ return UNALLOCATED_ENTRY;
}
}
}
@@ -676,30 +887,50 @@
}
size_t OopStorage::block_count() const {
- return _block_count;
+ WithActiveArray wab(this);
+ // Count access is racy, but don't care.
+ return wab.active_array().block_count();
}
size_t OopStorage::total_memory_usage() const {
size_t total_size = sizeof(OopStorage);
total_size += strlen(name()) + 1;
- total_size += block_count() * Block::allocation_size();
+ total_size += sizeof(BlockArray);
+ WithActiveArray wab(this);
+ const BlockArray& blocks = wab.active_array();
+ // Count access is racy, but don't care.
+ total_size += blocks.block_count() * Block::allocation_size();
+ total_size += blocks.size() * sizeof(Block*);
return total_size;
}
// Parallel iteration support
-static char* not_started_marker_dummy = NULL;
-static void* const not_started_marker = ¬_started_marker_dummy;
+uint OopStorage::BasicParState::default_estimated_thread_count(bool concurrent) {
+ return concurrent ? ConcGCThreads : ParallelGCThreads;
+}
-OopStorage::BasicParState::BasicParState(OopStorage* storage, bool concurrent) :
+OopStorage::BasicParState::BasicParState(const OopStorage* storage,
+ uint estimated_thread_count,
+ bool concurrent) :
_storage(storage),
- _next_block(not_started_marker),
+ _active_array(_storage->obtain_active_array()),
+ _block_count(0), // initialized properly below
+ _next_block(0),
+ _estimated_thread_count(estimated_thread_count),
_concurrent(concurrent)
{
+ assert(estimated_thread_count > 0, "estimated thread count must be positive");
update_iteration_state(true);
+ // Get the block count *after* iteration state updated, so concurrent
+ // empty block deletion is suppressed and can't reduce the count. But
+ // ensure the count we use was written after the block with that count
+ // was fully initialized; see BlockArray::push.
+ _block_count = _active_array->block_count_acquire();
}
OopStorage::BasicParState::~BasicParState() {
+ _storage->relinquish_block_array(_active_array);
update_iteration_state(false);
}
@@ -711,29 +942,49 @@
}
}
-void OopStorage::BasicParState::ensure_iteration_started() {
- if (!_concurrent) {
- assert_at_safepoint();
+bool OopStorage::BasicParState::claim_next_segment(IterationData* data) {
+ data->_processed += data->_segment_end - data->_segment_start;
+ size_t start = OrderAccess::load_acquire(&_next_block);
+ if (start >= _block_count) {
+ return finish_iteration(data); // No more blocks available.
}
- assert(!_concurrent || _storage->_concurrent_iteration_active, "invariant");
- // Ensure _next_block is not the not_started_marker, setting it to
- // the _active_head to start the iteration if necessary.
- if (OrderAccess::load_acquire(&_next_block) == not_started_marker) {
- Atomic::cmpxchg(_storage->_active_head, &_next_block, not_started_marker);
+ // Try to claim several at a time, but not *too* many. We want to
+ // avoid deciding there are many available and selecting a large
+ // quantity, get delayed, and then end up claiming most or all of
+ // the remaining largish amount of work, leaving nothing for other
+ // threads to do. But too small a step can lead to contention
+ // over _next_block, esp. when the work per block is small.
+ size_t max_step = 10;
+ size_t remaining = _block_count - start;
+ size_t step = MIN2(max_step, 1 + (remaining / _estimated_thread_count));
+ // Atomic::add with possible overshoot. This can perform better
+ // than a CAS loop on some platforms when there is contention.
+ // We can cope with the uncertainty by recomputing start/end from
+ // the result of the add, and dealing with potential overshoot.
+ size_t end = Atomic::add(step, &_next_block);
+ // _next_block may have changed, so recompute start from result of add.
+ start = end - step;
+ // _next_block may have changed so much that end has overshot.
+ end = MIN2(end, _block_count);
+ // _next_block may have changed so much that even start has overshot.
+ if (start < _block_count) {
+ // Record claimed segment for iteration.
+ data->_segment_start = start;
+ data->_segment_end = end;
+ return true; // Success.
+ } else {
+ // No more blocks to claim.
+ return finish_iteration(data);
}
- assert(_next_block != not_started_marker, "postcondition");
}
-OopStorage::Block* OopStorage::BasicParState::claim_next_block() {
- assert(_next_block != not_started_marker, "Iteration not started");
- void* next = _next_block;
- while (next != NULL) {
- void* new_next = _storage->_active_list.next(*static_cast<Block*>(next));
- void* fetched = Atomic::cmpxchg(new_next, &_next_block, next);
- if (fetched == next) break; // Claimed.
- next = fetched;
- }
- return static_cast<Block*>(next);
+bool OopStorage::BasicParState::finish_iteration(const IterationData* data) const {
+ log_debug(oopstorage, blocks, stats)
+ ("Parallel iteration on %s: blocks = " SIZE_FORMAT
+ ", processed = " SIZE_FORMAT " (%2.f%%)",
+ _storage->name(), _block_count, data->_processed,
+ percent_of(data->_processed, _block_count));
+ return false;
}
const char* OopStorage::name() const { return _name; }
@@ -742,7 +993,7 @@
void OopStorage::print_on(outputStream* st) const {
size_t allocations = _allocation_count;
- size_t blocks = _block_count;
+ size_t blocks = _active_array->block_count();
double data_size = section_size * section_count;
double alloc_percentage = percent_of((double)allocations, blocks * data_size);
--- a/src/hotspot/share/gc/shared/oopStorage.hpp Thu May 03 14:13:20 2018 -0700
+++ b/src/hotspot/share/gc/shared/oopStorage.hpp Thu May 03 17:36:50 2018 -0400
@@ -170,27 +170,11 @@
// classes. C++03 introduced access for nested classes with DR45, but xlC
// version 12 rejects it.
NOT_AIX( private: )
- class Block; // Forward decl; defined in .inline.hpp file.
- class BlockList; // Forward decl for BlockEntry friend decl.
-
- class BlockEntry {
- friend class BlockList;
+ class Block; // Fixed-size array of oops, plus bookkeeping.
+ class BlockArray; // Array of Blocks, plus bookkeeping.
+ class BlockEntry; // Provides BlockList links in a Block.
- // Members are mutable, and we deal exclusively with pointers to
- // const, to make const blocks easier to use; a block being const
- // doesn't prevent modifying its list state.
- mutable const Block* _prev;
- mutable const Block* _next;
-
- // Noncopyable.
- BlockEntry(const BlockEntry&);
- BlockEntry& operator=(const BlockEntry&);
-
- public:
- BlockEntry();
- ~BlockEntry();
- };
-
+ // Doubly-linked list of Blocks.
class BlockList {
const Block* _head;
const Block* _tail;
@@ -205,6 +189,7 @@
~BlockList();
Block* head();
+ Block* tail();
const Block* chead() const;
const Block* ctail() const;
@@ -219,19 +204,34 @@
void unlink(const Block& block);
};
+ // RCU-inspired protection of access to _active_array.
+ class ProtectActive {
+ volatile uint _enter;
+ volatile uint _exit[2];
+
+ public:
+ ProtectActive();
+
+ uint read_enter();
+ void read_exit(uint enter_value);
+ void write_synchronize();
+ };
+
private:
const char* _name;
- BlockList _active_list;
+ BlockArray* _active_array;
BlockList _allocate_list;
- Block* volatile _active_head;
Block* volatile _deferred_updates;
Mutex* _allocate_mutex;
Mutex* _active_mutex;
- // Counts are volatile for racy unlocked accesses.
+ // Volatile for racy unlocked accesses.
volatile size_t _allocation_count;
- volatile size_t _block_count;
+
+ // Protection for _active_array.
+ mutable ProtectActive _protect_active;
+
// mutable because this gets set even for const iteration.
mutable bool _concurrent_iteration_active;
@@ -239,6 +239,13 @@
void delete_empty_block(const Block& block);
bool reduce_deferred_updates();
+ // Managing _active_array.
+ bool expand_active_array();
+ void replace_active_array(BlockArray* new_array);
+ BlockArray* obtain_active_array() const;
+ void relinquish_block_array(BlockArray* array) const;
+ class WithActiveArray; // RAII helper for active array access.
+
template<typename F, typename Storage>
static bool iterate_impl(F f, Storage* storage);
--- a/src/hotspot/share/gc/shared/oopStorage.inline.hpp Thu May 03 14:13:20 2018 -0700
+++ b/src/hotspot/share/gc/shared/oopStorage.inline.hpp Thu May 03 17:36:50 2018 -0400
@@ -30,10 +30,107 @@
#include "metaprogramming/isConst.hpp"
#include "oops/oop.hpp"
#include "runtime/safepoint.hpp"
+#include "utilities/align.hpp"
#include "utilities/count_trailing_zeros.hpp"
#include "utilities/debug.hpp"
#include "utilities/globalDefinitions.hpp"
+// Array of all active blocks. Refcounted for lock-free reclaim of
+// old array when a new array is allocated for expansion.
+class OopStorage::BlockArray {
+ friend class OopStorage::TestAccess;
+
+ size_t _size;
+ volatile size_t _block_count;
+ mutable volatile int _refcount;
+ // Block* _blocks[1]; // Pseudo flexible array member.
+
+ BlockArray(size_t size);
+ ~BlockArray();
+
+ // Noncopyable
+ BlockArray(const BlockArray&);
+ BlockArray& operator=(const BlockArray&);
+
+ static size_t blocks_offset();
+ Block* const* base_ptr() const;
+
+ Block* const* block_ptr(size_t index) const;
+ Block** block_ptr(size_t index);
+
+public:
+ static BlockArray* create(size_t size, AllocFailType alloc_fail = AllocFailStrategy::EXIT_OOM);
+ static void destroy(BlockArray* ba);
+
+ inline Block* at(size_t i) const;
+
+ size_t size() const;
+ size_t block_count() const;
+ size_t block_count_acquire() const;
+ void increment_refcount() const;
+ bool decrement_refcount() const; // Return true if zero, otherwise false
+
+ // Support for OopStorage::allocate.
+ // Add block to the end of the array. Updates block count at the
+ // end of the operation, with a release_store. Returns true if the
+ // block was added, false if there was no room available.
+ // precondition: owner's _allocation_mutex is locked, or at safepoint.
+ bool push(Block* block);
+
+ // Support OopStorage::delete_empty_blocks_xxx operations.
+ // Remove block from the array.
+ // precondition: block must be present at its active_index element.
+ void remove(Block* block);
+
+ void copy_from(const BlockArray* from);
+};
+
+inline size_t OopStorage::BlockArray::blocks_offset() {
+ return align_up(sizeof(BlockArray), sizeof(Block*));
+}
+
+inline OopStorage::Block* const* OopStorage::BlockArray::base_ptr() const {
+ const void* ptr = reinterpret_cast<const char*>(this) + blocks_offset();
+ return reinterpret_cast<Block* const*>(ptr);
+}
+
+inline OopStorage::Block* const* OopStorage::BlockArray::block_ptr(size_t index) const {
+ return base_ptr() + index;
+}
+
+inline OopStorage::Block** OopStorage::BlockArray::block_ptr(size_t index) {
+ return const_cast<Block**>(base_ptr() + index);
+}
+
+inline OopStorage::Block* OopStorage::BlockArray::at(size_t index) const {
+ assert(index < _block_count, "precondition");
+ return *block_ptr(index);
+}
+
+// A Block has an embedded BlockEntry to provide the links between
+// Blocks in a BlockList.
+class OopStorage::BlockEntry {
+ friend class OopStorage::BlockList;
+
+ // Members are mutable, and we deal exclusively with pointers to
+ // const, to make const blocks easier to use; a block being const
+ // doesn't prevent modifying its list state.
+ mutable const Block* _prev;
+ mutable const Block* _next;
+
+ // Noncopyable.
+ BlockEntry(const BlockEntry&);
+ BlockEntry& operator=(const BlockEntry&);
+
+public:
+ BlockEntry();
+ ~BlockEntry();
+};
+
+// Fixed-sized array of oops, plus bookkeeping data.
+// All blocks are in the storage's _active_array, at the block's _active_index.
+// Non-full blocks are in the storage's _allocate_list, linked through the
+// block's _allocate_entry. Empty blocks are at the end of that list.
class OopStorage::Block /* No base class, to avoid messing up alignment. */ {
// _data must be the first non-static data member, for alignment.
oop _data[BitsPerWord];
@@ -42,7 +139,7 @@
volatile uintx _allocated_bitmask; // One bit per _data element.
const OopStorage* _owner;
void* _memory; // Unaligned storage containing block.
- BlockEntry _active_entry;
+ size_t _active_index;
BlockEntry _allocate_entry;
Block* volatile _deferred_updates_next;
volatile uintx _release_refcount;
@@ -61,7 +158,6 @@
Block& operator=(const Block&);
public:
- static const BlockEntry& get_active_entry(const Block& block);
static const BlockEntry& get_allocate_entry(const Block& block);
static size_t allocation_size();
@@ -84,6 +180,10 @@
bool contains(const oop* ptr) const;
+ size_t active_index() const;
+ void set_active_index(size_t index);
+ static size_t active_index_safe(const Block* block); // Returns 0 if access fails.
+
// Returns NULL if ptr is not in a block or not allocated in that block.
static Block* block_for_ptr(const OopStorage* owner, const oop* ptr);
@@ -101,6 +201,10 @@
return const_cast<Block*>(_head);
}
+inline OopStorage::Block* OopStorage::BlockList::tail() {
+ return const_cast<Block*>(_tail);
+}
+
inline const OopStorage::Block* OopStorage::BlockList::chead() const {
return _head;
}
@@ -253,9 +357,10 @@
// Propagate const/non-const iteration to the block layer, by using
// const or non-const blocks as corresponding to Storage.
typedef typename Conditional<IsConst<Storage>::value, const Block*, Block*>::type BlockPtr;
- for (BlockPtr block = storage->_active_head;
- block != NULL;
- block = storage->_active_list.next(*block)) {
+ BlockArray* blocks = storage->_active_array;
+ size_t limit = blocks->block_count();
+ for (size_t i = 0; i < limit; ++i) {
+ BlockPtr block = blocks->at(i);
if (!block->iterate(f)) {
return false;
}
--- a/src/hotspot/share/gc/shared/oopStorageParState.hpp Thu May 03 14:13:20 2018 -0700
+++ b/src/hotspot/share/gc/shared/oopStorageParState.hpp Thu May 03 17:36:50 2018 -0400
@@ -36,9 +36,8 @@
//
// Concurrent Iteration
//
-// Iteration involves the _active_list, which contains all of the blocks owned
-// by a storage object. This is a doubly-linked list, linked through
-// dedicated fields in the blocks.
+// Iteration involves the _active_array (a BlockArray), which contains all of
+// the blocks owned by a storage object.
//
// At most one concurrent ParState can exist at a time for a given storage
// object.
@@ -48,27 +47,29 @@
// sets it false when the state is destroyed. These assignments are made with
// _active_mutex locked. Meanwhile, empty block deletion is not done while
// _concurrent_iteration_active is true. The flag check and the dependent
-// removal of a block from the _active_list is performed with _active_mutex
+// removal of a block from the _active_array is performed with _active_mutex
// locked. This prevents concurrent iteration and empty block deletion from
// interfering with with each other.
//
// Both allocate() and delete_empty_blocks_concurrent() lock the
-// _allocate_mutex while performing their respective list manipulations,
-// preventing them from interfering with each other.
+// _allocate_mutex while performing their respective list and array
+// manipulations, preventing them from interfering with each other.
//
-// When allocate() creates a new block, it is added to the front of the
-// _active_list. Then _active_head is set to the new block. When concurrent
-// iteration is started (by a parallel worker thread calling the state's
-// iterate() function), the current _active_head is used as the initial block
-// for the iteration, with iteration proceeding down the list headed by that
-// block.
+// When allocate() creates a new block, it is added to the end of the
+// _active_array. Then _active_array's _block_count is incremented to account
+// for the new block. When concurrent iteration is started (by a parallel
+// worker thread calling the state's iterate() function), the current
+// _active_array and its _block_count are captured for use by the iteration,
+// with iteration processing all blocks in that array up to that block count.
//
-// As a result, the list over which concurrent iteration operates is stable.
-// However, once the iteration is started, later allocations may add blocks to
-// the front of the list that won't be examined by the iteration. And while
-// the list is stable, concurrent allocate() and release() operations may
-// change the set of allocated entries in a block at any time during the
-// iteration.
+// As a result, the sequence over which concurrent iteration operates is
+// stable. However, once the iteration is started, later allocations may add
+// blocks to the end of the array that won't be examined by the iteration.
+// An allocation may even require expansion of the array, so the iteration is
+// no longer processing the current array, but rather the previous one.
+// And while the sequence is stable, concurrent allocate() and release()
+// operations may change the set of allocated entries in a block at any time
+// during the iteration.
//
// As a result, a concurrent iteration handler must accept that some
// allocations and releases that occur after the iteration started will not be
@@ -138,36 +139,49 @@
// invoked on p.
class OopStorage::BasicParState {
- OopStorage* _storage;
- void* volatile _next_block;
+ const OopStorage* _storage;
+ BlockArray* _active_array;
+ size_t _block_count;
+ volatile size_t _next_block;
+ uint _estimated_thread_count;
bool _concurrent;
// Noncopyable.
BasicParState(const BasicParState&);
BasicParState& operator=(const BasicParState&);
+ struct IterationData;
+
void update_iteration_state(bool value);
- void ensure_iteration_started();
- Block* claim_next_block();
+ bool claim_next_segment(IterationData* data);
+ bool finish_iteration(const IterationData* data) const;
// Wrapper for iteration handler; ignore handler result and return true.
template<typename F> class AlwaysTrueFn;
public:
- BasicParState(OopStorage* storage, bool concurrent);
+ BasicParState(const OopStorage* storage,
+ uint estimated_thread_count,
+ bool concurrent);
~BasicParState();
template<bool is_const, typename F> void iterate(F f);
+
+ static uint default_estimated_thread_count(bool concurrent);
};
template<bool concurrent, bool is_const>
class OopStorage::ParState {
BasicParState _basic_state;
+ typedef typename Conditional<is_const,
+ const OopStorage*,
+ OopStorage*>::type StoragePtr;
+
public:
- ParState(const OopStorage* storage) :
- // For simplicity, always recorded as non-const.
- _basic_state(const_cast<OopStorage*>(storage), concurrent)
+ ParState(StoragePtr storage,
+ uint estimated_thread_count = BasicParState::default_estimated_thread_count(concurrent)) :
+ _basic_state(storage, estimated_thread_count, concurrent)
{}
template<typename F> void iterate(F f);
@@ -179,8 +193,9 @@
BasicParState _basic_state;
public:
- ParState(OopStorage* storage) :
- _basic_state(storage, false)
+ ParState(OopStorage* storage,
+ uint estimated_thread_count = BasicParState::default_estimated_thread_count(false)) :
+ _basic_state(storage, estimated_thread_count, false)
{}
template<typename F> void iterate(F f);
--- a/src/hotspot/share/gc/shared/oopStorageParState.inline.hpp Thu May 03 14:13:20 2018 -0700
+++ b/src/hotspot/share/gc/shared/oopStorageParState.inline.hpp Thu May 03 17:36:50 2018 -0400
@@ -41,14 +41,26 @@
bool operator()(OopPtr ptr) const { _f(ptr); return true; }
};
+struct OopStorage::BasicParState::IterationData {
+ size_t _segment_start;
+ size_t _segment_end;
+ size_t _processed;
+};
+
template<bool is_const, typename F>
inline void OopStorage::BasicParState::iterate(F f) {
// Wrap f in ATF so we can use Block::iterate.
AlwaysTrueFn<F> atf_f(f);
- ensure_iteration_started();
- typename Conditional<is_const, const Block*, Block*>::type block;
- while ((block = claim_next_block()) != NULL) {
- block->iterate(atf_f);
+ IterationData data = {}; // zero initialize.
+ while (claim_next_segment(&data)) {
+ assert(data._segment_start < data._segment_end, "invariant");
+ assert(data._segment_end <= _block_count, "invariant");
+ typedef typename Conditional<is_const, const Block*, Block*>::type BlockPtr;
+ size_t i = data._segment_start;
+ do {
+ BlockPtr block = _active_array->at(i);
+ block->iterate(atf_f);
+ } while (++i < data._segment_end);
}
}
--- a/test/hotspot/gtest/gc/shared/test_oopStorage.cpp Thu May 03 14:13:20 2018 -0700
+++ b/test/hotspot/gtest/gc/shared/test_oopStorage.cpp Thu May 03 17:36:50 2018 -0400
@@ -53,9 +53,10 @@
public:
typedef OopStorage::Block Block;
typedef OopStorage::BlockList BlockList;
+ typedef OopStorage::BlockArray BlockArray;
- static BlockList& active_list(OopStorage& storage) {
- return storage._active_list;
+ static BlockArray& active_array(const OopStorage& storage) {
+ return *storage._active_array;
}
static BlockList& allocate_list(OopStorage& storage) {
@@ -96,20 +97,25 @@
static size_t memory_per_block() {
return Block::allocation_size();
}
+
+ static void block_array_set_block_count(BlockArray* blocks, size_t count) {
+ blocks->_block_count = count;
+ }
};
typedef OopStorage::TestAccess TestAccess;
-// --- FIXME: Should be just Block, but that collides with opto Block
-// when building with precompiled headers. There really should be
-// an opto namespace.
+
+// The "Oop" prefix is to avoid collision with similar opto names when
+// building with precompiled headers, or for consistency with that
+// workaround. There really should be an opto namespace.
typedef TestAccess::Block OopBlock;
-// --- FIXME: Similarly, this typedef collides with opto BlockList.
-// typedef TestAccess::BlockList BlockList;
+typedef TestAccess::BlockList OopBlockList;
+typedef TestAccess::BlockArray OopBlockArray;
// Using EXPECT_EQ can't use NULL directly. Otherwise AIX build breaks.
const OopBlock* const NULL_BLOCK = NULL;
-static size_t list_length(const TestAccess::BlockList& list) {
+static size_t list_length(const OopBlockList& list) {
size_t result = 0;
for (const OopBlock* block = list.chead();
block != NULL;
@@ -119,7 +125,7 @@
return result;
}
-static void clear_list(TestAccess::BlockList& list) {
+static void clear_list(OopBlockList& list) {
OopBlock* next;
for (OopBlock* block = list.head(); block != NULL; block = next) {
next = list.next(*block);
@@ -127,7 +133,7 @@
}
}
-static bool is_list_empty(const TestAccess::BlockList& list) {
+static bool is_list_empty(const OopBlockList& list) {
return list.chead() == NULL;
}
@@ -149,7 +155,7 @@
}
static size_t empty_block_count(const OopStorage& storage) {
- const TestAccess::BlockList& list = TestAccess::allocate_list(storage);
+ const OopBlockList& list = TestAccess::allocate_list(storage);
size_t count = 0;
for (const OopBlock* block = list.ctail();
(block != NULL) && block->is_empty();
@@ -158,6 +164,20 @@
return count;
}
+static size_t active_count(const OopStorage& storage) {
+ return TestAccess::active_array(storage).block_count();
+}
+
+static OopBlock* active_head(const OopStorage& storage) {
+ OopBlockArray& ba = TestAccess::active_array(storage);
+ size_t count = ba.block_count();
+ if (count == 0) {
+ return NULL;
+ } else {
+ return ba.at(count - 1);
+ }
+}
+
class OopStorageTest : public ::testing::Test {
public:
OopStorageTest();
@@ -188,7 +208,6 @@
OopStorageTest::~OopStorageTest() {
clear_list(TestAccess::allocate_list(_storage));
- clear_list(TestAccess::active_list(_storage));
}
class OopStorageTestWithAllocation : public OopStorageTest {
@@ -227,7 +246,7 @@
static bool is_allocate_list_sorted(const OopStorage& storage) {
// The allocate_list isn't strictly sorted. Rather, all empty
// blocks are segregated to the end of the list.
- const TestAccess::BlockList& list = TestAccess::allocate_list(storage);
+ const OopBlockList& list = TestAccess::allocate_list(storage);
const OopBlock* block = list.ctail();
for ( ; (block != NULL) && block->is_empty(); block = list.prev(*block)) {}
for ( ; block != NULL; block = list.prev(*block)) {
@@ -238,25 +257,25 @@
return true;
}
-static size_t total_allocation_count(const TestAccess::BlockList& list) {
+static size_t total_allocation_count(const OopStorage& storage) {
size_t total_count = 0;
- for (const OopBlock* block = list.chead();
- block != NULL;
- block = list.next(*block)) {
- total_count += TestAccess::block_allocation_count(*block);
+ const OopBlockArray& ba = TestAccess::active_array(storage);
+ size_t limit = active_count(storage);
+ for (size_t i = 0; i < limit; ++i) {
+ total_count += TestAccess::block_allocation_count(*ba.at(i));
}
return total_count;
}
TEST_VM_F(OopStorageTest, allocate_one) {
- EXPECT_TRUE(is_list_empty(TestAccess::active_list(_storage)));
+ EXPECT_EQ(0u, active_count(_storage));
EXPECT_TRUE(is_list_empty(TestAccess::allocate_list(_storage)));
oop* ptr = _storage.allocate();
EXPECT_TRUE(ptr != NULL);
EXPECT_EQ(1u, _storage.allocation_count());
- EXPECT_EQ(1u, list_length(TestAccess::active_list(_storage)));
+ EXPECT_EQ(1u, active_count(_storage));
EXPECT_EQ(1u, _storage.block_count());
EXPECT_EQ(1u, list_length(TestAccess::allocate_list(_storage)));
@@ -264,7 +283,7 @@
const OopBlock* block = TestAccess::allocate_list(_storage).chead();
EXPECT_NE(block, (OopBlock*)NULL);
- EXPECT_EQ(block, (TestAccess::active_list(_storage).chead()));
+ EXPECT_EQ(block, active_head(_storage));
EXPECT_FALSE(TestAccess::block_is_empty(*block));
EXPECT_FALSE(TestAccess::block_is_full(*block));
EXPECT_EQ(1u, TestAccess::block_allocation_count(*block));
@@ -272,7 +291,7 @@
release_entry(_storage, ptr);
EXPECT_EQ(0u, _storage.allocation_count());
- EXPECT_EQ(1u, list_length(TestAccess::active_list(_storage)));
+ EXPECT_EQ(1u, active_count(_storage));
EXPECT_EQ(1u, _storage.block_count());
EXPECT_EQ(1u, list_length(TestAccess::allocate_list(_storage)));
@@ -280,7 +299,7 @@
const OopBlock* new_block = TestAccess::allocate_list(_storage).chead();
EXPECT_EQ(block, new_block);
- EXPECT_EQ(block, (TestAccess::active_list(_storage).chead()));
+ EXPECT_EQ(block, active_head(_storage));
EXPECT_TRUE(TestAccess::block_is_empty(*block));
EXPECT_FALSE(TestAccess::block_is_full(*block));
EXPECT_EQ(0u, TestAccess::block_allocation_count(*block));
@@ -290,20 +309,19 @@
static const size_t max_entries = 1000;
oop* entries[max_entries];
- TestAccess::BlockList& active_list = TestAccess::active_list(_storage);
- TestAccess::BlockList& allocate_list = TestAccess::allocate_list(_storage);
+ OopBlockList& allocate_list = TestAccess::allocate_list(_storage);
- EXPECT_TRUE(is_list_empty(active_list));
+ EXPECT_EQ(0u, active_count(_storage));
EXPECT_EQ(0u, _storage.block_count());
EXPECT_TRUE(is_list_empty(allocate_list));
size_t allocated = 0;
for ( ; allocated < max_entries; ++allocated) {
EXPECT_EQ(allocated, _storage.allocation_count());
- if (!is_list_empty(active_list)) {
- EXPECT_EQ(1u, list_length(active_list));
+ if (active_count(_storage) != 0) {
+ EXPECT_EQ(1u, active_count(_storage));
EXPECT_EQ(1u, _storage.block_count());
- const OopBlock& block = *active_list.chead();
+ const OopBlock& block = *TestAccess::active_array(_storage).at(0);
EXPECT_EQ(allocated, TestAccess::block_allocation_count(block));
if (TestAccess::block_is_full(block)) {
break;
@@ -316,10 +334,10 @@
}
EXPECT_EQ(allocated, _storage.allocation_count());
- EXPECT_EQ(1u, list_length(active_list));
+ EXPECT_EQ(1u, active_count(_storage));
EXPECT_EQ(1u, _storage.block_count());
EXPECT_TRUE(is_list_empty(allocate_list));
- const OopBlock& block = *active_list.chead();
+ const OopBlock& block = *TestAccess::active_array(_storage).at(0);
EXPECT_TRUE(TestAccess::block_is_full(block));
EXPECT_EQ(allocated, TestAccess::block_allocation_count(block));
@@ -336,19 +354,18 @@
static const size_t max_entries = 1000;
oop* entries[max_entries];
- TestAccess::BlockList& active_list = TestAccess::active_list(_storage);
- TestAccess::BlockList& allocate_list = TestAccess::allocate_list(_storage);
+ OopBlockList& allocate_list = TestAccess::allocate_list(_storage);
EXPECT_EQ(0u, empty_block_count(_storage));
entries[0] = _storage.allocate();
ASSERT_TRUE(entries[0] != NULL);
- EXPECT_EQ(1u, list_length(active_list));
+ EXPECT_EQ(1u, active_count(_storage));
EXPECT_EQ(1u, _storage.block_count());
EXPECT_EQ(1u, list_length(allocate_list));
EXPECT_EQ(0u, empty_block_count(_storage));
- const OopBlock* block = active_list.chead();
+ const OopBlock* block = TestAccess::active_array(_storage).at(0);
EXPECT_EQ(1u, TestAccess::block_allocation_count(*block));
EXPECT_EQ(block, allocate_list.chead());
@@ -363,14 +380,14 @@
EXPECT_EQ(1u, list_length(allocate_list));
block = allocate_list.chead();
EXPECT_EQ(1u, TestAccess::block_allocation_count(*block));
- EXPECT_EQ(block, active_list.chead());
+ EXPECT_EQ(block, active_head(_storage));
} else if (TestAccess::block_is_full(*block)) {
EXPECT_TRUE(is_list_empty(allocate_list));
block = NULL;
} else {
EXPECT_FALSE(is_list_empty(allocate_list));
EXPECT_EQ(block, allocate_list.chead());
- EXPECT_EQ(block, active_list.chead());
+ EXPECT_EQ(block, active_head(_storage));
}
}
@@ -378,20 +395,18 @@
EXPECT_NE(0u, TestAccess::block_allocation_count(*block));
EXPECT_FALSE(is_list_empty(allocate_list));
EXPECT_EQ(block, allocate_list.chead());
- EXPECT_EQ(block, active_list.chead());
+ EXPECT_EQ(block, active_head(_storage));
}
- size_t active_count = list_length(active_list);
-
for (size_t i = 0; i < max_entries; ++i) {
release_entry(_storage, entries[i]);
EXPECT_TRUE(is_allocate_list_sorted(_storage));
- EXPECT_EQ(max_entries - (i + 1), total_allocation_count(active_list));
+ EXPECT_EQ(max_entries - (i + 1), total_allocation_count(_storage));
}
- EXPECT_EQ(list_length(active_list), list_length(allocate_list));
- EXPECT_EQ(list_length(active_list), _storage.block_count());
- EXPECT_EQ(list_length(active_list), empty_block_count(_storage));
+ EXPECT_EQ(active_count(_storage), list_length(allocate_list));
+ EXPECT_EQ(active_count(_storage), _storage.block_count());
+ EXPECT_EQ(active_count(_storage), empty_block_count(_storage));
for (const OopBlock* block = allocate_list.chead();
block != NULL;
block = allocate_list.next(*block)) {
@@ -405,10 +420,9 @@
EXPECT_EQ(0u, empty_block_count(_storage));
- TestAccess::BlockList& active_list = TestAccess::active_list(_storage);
- TestAccess::BlockList& allocate_list = TestAccess::allocate_list(_storage);
+ OopBlockList& allocate_list = TestAccess::allocate_list(_storage);
- EXPECT_EQ(_max_entries, total_allocation_count(active_list));
+ EXPECT_EQ(_max_entries, total_allocation_count(_storage));
EXPECT_GE(1u, list_length(allocate_list));
// Release all entries in "random" order.
@@ -418,14 +432,14 @@
release_entry(_storage, _entries[i]);
_entries[i] = NULL;
++released;
- EXPECT_EQ(_max_entries - released, total_allocation_count(active_list));
+ EXPECT_EQ(_max_entries - released, total_allocation_count(_storage));
EXPECT_TRUE(is_allocate_list_sorted(_storage));
}
}
- EXPECT_EQ(list_length(active_list), list_length(allocate_list));
- EXPECT_EQ(list_length(active_list), _storage.block_count());
- EXPECT_EQ(0u, total_allocation_count(active_list));
+ EXPECT_EQ(active_count(_storage), list_length(allocate_list));
+ EXPECT_EQ(active_count(_storage), _storage.block_count());
+ EXPECT_EQ(0u, total_allocation_count(_storage));
EXPECT_EQ(list_length(allocate_list), empty_block_count(_storage));
}
@@ -436,10 +450,9 @@
EXPECT_EQ(0u, empty_block_count(_storage));
- TestAccess::BlockList& active_list = TestAccess::active_list(_storage);
- TestAccess::BlockList& allocate_list = TestAccess::allocate_list(_storage);
+ OopBlockList& allocate_list = TestAccess::allocate_list(_storage);
- EXPECT_EQ(_max_entries, total_allocation_count(active_list));
+ EXPECT_EQ(_max_entries, total_allocation_count(_storage));
EXPECT_GE(1u, list_length(allocate_list));
// Release all entries in "random" order, "randomly" interspersed
@@ -452,20 +465,20 @@
_entries[i] = NULL;
++released;
++total_released;
- EXPECT_EQ(_max_entries - released, total_allocation_count(active_list));
+ EXPECT_EQ(_max_entries - released, total_allocation_count(_storage));
EXPECT_TRUE(is_allocate_list_sorted(_storage));
if (total_released % allocate_step == 0) {
_entries[i] = _storage.allocate();
--released;
- EXPECT_EQ(_max_entries - released, total_allocation_count(active_list));
+ EXPECT_EQ(_max_entries - released, total_allocation_count(_storage));
EXPECT_TRUE(is_allocate_list_sorted(_storage));
}
}
}
- EXPECT_EQ(list_length(active_list), list_length(allocate_list));
- EXPECT_EQ(list_length(active_list), _storage.block_count());
- EXPECT_EQ(0u, total_allocation_count(active_list));
+ EXPECT_EQ(active_count(_storage), list_length(allocate_list));
+ EXPECT_EQ(active_count(_storage), _storage.block_count());
+ EXPECT_EQ(0u, total_allocation_count(_storage));
EXPECT_EQ(list_length(allocate_list), empty_block_count(_storage));
}
@@ -1015,9 +1028,7 @@
}
TEST_VM_F(OopStorageTestWithAllocation, delete_empty_blocks_safepoint) {
- TestAccess::BlockList& active_list = TestAccess::active_list(_storage);
-
- size_t initial_active_size = list_length(active_list);
+ size_t initial_active_size = active_count(_storage);
EXPECT_EQ(initial_active_size, _storage.block_count());
ASSERT_LE(3u, initial_active_size); // Need at least 3 blocks for test
@@ -1026,7 +1037,7 @@
release_entry(_storage, _entries[i]);
}
- EXPECT_EQ(initial_active_size, list_length(active_list));
+ EXPECT_EQ(initial_active_size, active_count(_storage));
EXPECT_EQ(initial_active_size, _storage.block_count());
EXPECT_EQ(3u, empty_block_count(_storage));
@@ -1036,14 +1047,12 @@
VMThread::execute(&op);
}
EXPECT_EQ(0u, empty_block_count(_storage));
- EXPECT_EQ(initial_active_size - 3, list_length(active_list));
+ EXPECT_EQ(initial_active_size - 3, active_count(_storage));
EXPECT_EQ(initial_active_size - 3, _storage.block_count());
}
TEST_VM_F(OopStorageTestWithAllocation, delete_empty_blocks_concurrent) {
- TestAccess::BlockList& active_list = TestAccess::active_list(_storage);
-
- size_t initial_active_size = list_length(active_list);
+ size_t initial_active_size = active_count(_storage);
EXPECT_EQ(initial_active_size, _storage.block_count());
ASSERT_LE(3u, initial_active_size); // Need at least 3 blocks for test
@@ -1052,13 +1061,13 @@
release_entry(_storage, _entries[i]);
}
- EXPECT_EQ(initial_active_size, list_length(active_list));
+ EXPECT_EQ(initial_active_size, active_count(_storage));
EXPECT_EQ(initial_active_size, _storage.block_count());
EXPECT_EQ(3u, empty_block_count(_storage));
_storage.delete_empty_blocks_concurrent();
EXPECT_EQ(0u, empty_block_count(_storage));
- EXPECT_EQ(initial_active_size - 3, list_length(active_list));
+ EXPECT_EQ(initial_active_size - 3, active_count(_storage));
EXPECT_EQ(initial_active_size - 3, _storage.block_count());
}
@@ -1161,23 +1170,21 @@
#endif // !PRODUCT
-//////////////////////////////////////////////////////////////////////////////
-// Unit tests for block lists
-
-class OopStorageBlockListTest : public ::testing::Test {
-public:
- OopStorageBlockListTest() {
+class OopStorageBlockCollectionTest : public ::testing::Test {
+protected:
+ OopStorageBlockCollectionTest() {
for (size_t i = 0; i < nvalues; ++i) {
values[i] = OopBlock::new_block(pseudo_owner());
}
}
- ~OopStorageBlockListTest() {
+ ~OopStorageBlockCollectionTest() {
for (size_t i = 0; i < nvalues; ++i) {
OopBlock::delete_block(*values[i]);
}
}
+public:
static const size_t nvalues = 10;
OopBlock* values[nvalues];
@@ -1190,11 +1197,13 @@
}
};
-const size_t OopStorageBlockListTest::nvalues;
-const void* const OopStorageBlockListTest::_pseudo_owner[] = {};
+const size_t OopStorageBlockCollectionTest::nvalues;
+const void* const OopStorageBlockCollectionTest::_pseudo_owner[] = {};
+
+class OopStorageBlockListTest : public OopStorageBlockCollectionTest {};
TEST_F(OopStorageBlockListTest, empty_list) {
- TestAccess::BlockList list(&OopBlock::get_active_entry);
+ OopBlockList list(&OopBlock::get_allocate_entry);
EXPECT_TRUE(is_list_empty(list));
EXPECT_EQ(NULL_BLOCK, list.head());
@@ -1203,7 +1212,7 @@
}
TEST_F(OopStorageBlockListTest, push_back) {
- TestAccess::BlockList list(&OopBlock::get_active_entry);
+ OopBlockList list(&OopBlock::get_allocate_entry);
for (size_t i = 0; i < nvalues; ++i) {
list.push_back(*values[i]);
@@ -1233,7 +1242,7 @@
}
TEST_F(OopStorageBlockListTest, push_front) {
- TestAccess::BlockList list(&OopBlock::get_active_entry);
+ OopBlockList list(&OopBlock::get_allocate_entry);
for (size_t i = 0; i < nvalues; ++i) {
list.push_front(*values[i]);
@@ -1264,7 +1273,7 @@
class OopStorageBlockListTestWithList : public OopStorageBlockListTest {
public:
- OopStorageBlockListTestWithList() : list(&OopBlock::get_active_entry) {
+ OopStorageBlockListTestWithList() : list(&OopBlock::get_allocate_entry) {
for (size_t i = 0; i < nvalues; ++i) {
list.push_back(*values[i]);
}
@@ -1274,7 +1283,7 @@
clear_list(list);
}
- TestAccess::BlockList list;
+ OopBlockList list;
};
TEST_F(OopStorageBlockListTestWithList, unlink_front) {
@@ -1336,7 +1345,7 @@
}
TEST_F(OopStorageBlockListTest, single) {
- TestAccess::BlockList list(&OopBlock::get_active_entry);
+ OopBlockList list(&OopBlock::get_allocate_entry);
list.push_back(*values[0]);
EXPECT_EQ(NULL_BLOCK, list.next(*values[0]));
@@ -1351,31 +1360,79 @@
EXPECT_EQ(NULL_BLOCK, list.ctail());
}
-TEST_F(OopStorageBlockListTestWithList, two_lists) {
- TestAccess::BlockList list2(&OopBlock::get_allocate_entry);
- for (size_t i = 0; i < nvalues; ++i) {
- list2.push_front(*values[i]);
+class OopStorageBlockArrayTest : public OopStorageBlockCollectionTest {};
+
+TEST_F(OopStorageBlockArrayTest, empty_array) {
+ OopBlockArray* a = OopBlockArray::create(nvalues);
+
+ EXPECT_EQ(nvalues, a->size());
+ EXPECT_EQ(0u, a->block_count_acquire());
+ TestAccess::block_array_set_block_count(a, 2);
+ EXPECT_EQ(2u, a->block_count_acquire());
+ TestAccess::block_array_set_block_count(a, 0);
+ a->increment_refcount();
+ a->increment_refcount();
+ EXPECT_FALSE(a->decrement_refcount());
+ EXPECT_TRUE(a->decrement_refcount());
+
+ OopBlockArray::destroy(a);
+}
+
+TEST_F(OopStorageBlockArrayTest, push) {
+ OopBlockArray* a = OopBlockArray::create(nvalues - 1);
+
+ for (size_t i = 0; i < nvalues - 1; ++i) {
+ EXPECT_TRUE(a->push(values[i]));
+ EXPECT_EQ(i + 1, a->block_count_acquire());
+ EXPECT_EQ(values[i], a->at(i));
+ }
+ EXPECT_FALSE(a->push(values[nvalues - 1]));
+
+ TestAccess::block_array_set_block_count(a, 0);
+ OopBlockArray::destroy(a);
+}
+
+class OopStorageBlockArrayTestWithArray : public OopStorageBlockArrayTest {
+public:
+ OopStorageBlockArrayTestWithArray() : a(OopBlockArray::create(nvalues)) {
+ for (size_t i = 0; i < nvalues; ++i) {
+ a->push(values[i]);
+ }
}
- const OopBlock* active_block = list.chead();
- const OopBlock* allocate_block = list2.ctail();
- for (size_t i = 0; i < nvalues; ++i) {
- EXPECT_EQ(active_block, allocate_block);
- active_block = list.next(*active_block);
- allocate_block = list2.prev(*allocate_block);
+ ~OopStorageBlockArrayTestWithArray() {
+ TestAccess::block_array_set_block_count(a, 0);
+ OopBlockArray::destroy(a);
}
- EXPECT_EQ(NULL_BLOCK, active_block);
- EXPECT_EQ(NULL_BLOCK, allocate_block);
+
+ OopBlockArray* a;
+};
+
+TEST_F(OopStorageBlockArrayTestWithArray, remove0) {
+ a->remove(values[0]);
+ EXPECT_EQ(nvalues - 1, a->block_count_acquire());
+ EXPECT_EQ(values[nvalues - 1], a->at(0));
+ for (size_t i = 1; i < nvalues - 1; ++i) {
+ EXPECT_EQ(values[i], a->at(i));
+ }
+}
- for (size_t i = 0; i < nvalues; ++i) {
- list2.unlink(*values[i]);
+TEST_F(OopStorageBlockArrayTestWithArray, remove3) {
+ a->remove(values[3]);
+ EXPECT_EQ(nvalues - 1, a->block_count_acquire());
+ for (size_t i = 0; i < 3; ++i) {
+ EXPECT_EQ(values[i], a->at(i));
}
- EXPECT_TRUE(is_list_empty(list2));
+ EXPECT_EQ(values[nvalues - 1], a->at(3));
+ for (size_t i = 4; i < nvalues - 1; ++i) {
+ EXPECT_EQ(values[i], a->at(i));
+ }
+}
- active_block = list.chead();
- for (size_t i = 0; i < nvalues; ++i) {
- EXPECT_EQ(active_block, values[i]);
- active_block = list.next(*active_block);
+TEST_F(OopStorageBlockArrayTestWithArray, remove_last) {
+ a->remove(values[nvalues - 1]);
+ EXPECT_EQ(nvalues - 1, a->block_count_acquire());
+ for (size_t i = 0; i < nvalues - 1; ++i) {
+ EXPECT_EQ(values[i], a->at(i));
}
- EXPECT_EQ(NULL_BLOCK, active_block);
}
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/test/hotspot/gtest/gc/shared/test_oopStorage_parperf.cpp Thu May 03 17:36:50 2018 -0400
@@ -0,0 +1,233 @@
+/*
+ * Copyright (c) 2018, Oracle and/or its affiliates. All rights reserved.
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This code is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 only, as
+ * published by the Free Software Foundation.
+ *
+ * This code is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ * version 2 for more details (a copy is included in the LICENSE file that
+ * accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License version
+ * 2 along with this work; if not, write to the Free Software Foundation,
+ * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
+ * or visit www.oracle.com if you need additional information or have any
+ * questions.
+ */
+
+#include "precompiled.hpp"
+#include "gc/shared/oopStorage.inline.hpp"
+#include "gc/shared/oopStorageParState.inline.hpp"
+#include "gc/shared/workgroup.hpp"
+#include "logging/log.hpp"
+#include "logging/logConfiguration.hpp"
+#include "memory/allocation.inline.hpp"
+#include "memory/iterator.inline.hpp"
+#include "runtime/interfaceSupport.inline.hpp"
+#include "runtime/os.hpp"
+#include "runtime/thread.hpp"
+#include "runtime/vm_operations.hpp"
+#include "runtime/vmThread.hpp"
+#include "utilities/debug.hpp"
+#include "utilities/ostream.hpp"
+#include "utilities/ticks.inline.hpp"
+
+#include "unittest.hpp"
+
+// This "test" doesn't really verify much. Rather, it's mostly a
+// microbenchmark for OopStorage parallel iteration. It executes
+// parallel iteration with varying numbers of threads on an storage
+// object containing a large number of entries, and logs some stats
+// about the distribution and performance of the iteration.
+
+// Parallel iteration not available unless INCLUDE_ALL_GCS
+#if INCLUDE_ALL_GCS
+
+const uint _max_workers = 10;
+static uint _num_workers = 0;
+const size_t _storage_entries = 1000000;
+
+class OopStorageParIterPerf : public ::testing::Test {
+public:
+ OopStorageParIterPerf();
+ ~OopStorageParIterPerf();
+
+ WorkGang* workers() const;
+
+ class VM_ParStateTime;
+ class Task;
+ class Closure;
+
+ Tickspan run_task(Task* task, uint nthreads);
+ void show_task(const Task* task, Tickspan duration, uint nthreads);
+ void run_test(uint nthreads);
+
+ static WorkGang* _workers;
+
+ static const int _active_rank = Mutex::leaf - 1;
+ static const int _allocate_rank = Mutex::leaf;
+
+ Mutex _allocate_mutex;
+ Mutex _active_mutex;
+ OopStorage _storage;
+ oop* _entries[_storage_entries];
+};
+
+WorkGang* OopStorageParIterPerf::_workers = NULL;
+
+WorkGang* OopStorageParIterPerf::workers() const {
+ if (_workers == NULL) {
+ WorkGang* wg = new WorkGang("OopStorageParIterPerf workers",
+ _num_workers,
+ false,
+ false);
+ wg->initialize_workers();
+ wg->update_active_workers(_num_workers);
+ _workers = wg;
+ }
+ return _workers;
+}
+
+OopStorageParIterPerf::OopStorageParIterPerf() :
+ _allocate_mutex(_allocate_rank,
+ "test_OopStorage_parperf_allocate",
+ false,
+ Mutex::_safepoint_check_never),
+ _active_mutex(_active_rank,
+ "test_OopStorage_parperf_active",
+ false,
+ Mutex::_safepoint_check_never),
+ _storage("Test Storage", &_allocate_mutex, &_active_mutex)
+{
+ for (size_t i = 0; i < _storage_entries; ++i) {
+ _entries[i] = _storage.allocate();
+ }
+ _num_workers = MIN2(_max_workers, (uint)os::processor_count());
+}
+
+OopStorageParIterPerf::~OopStorageParIterPerf() {
+ _storage.release(_entries, ARRAY_SIZE(_entries));
+}
+
+class OopStorageParIterPerf::VM_ParStateTime : public VM_GTestExecuteAtSafepoint {
+public:
+ VM_ParStateTime(WorkGang* workers, AbstractGangTask* task, uint nthreads) :
+ _workers(workers), _task(task), _nthreads(nthreads)
+ {}
+
+ void doit() {
+ _workers->run_task(_task, _nthreads);
+ }
+
+private:
+ WorkGang* _workers;
+ AbstractGangTask* _task;
+ uint _nthreads;
+};
+
+class OopStorageParIterPerf::Task : public AbstractGangTask {
+ typedef OopStorage::ParState<false, false> StateType;
+
+ Tickspan* _worker_times;
+ StateType _state;
+ OopClosure* _closure;
+
+public:
+ Task(OopStorage* storage, OopClosure* closure, uint nthreads) :
+ AbstractGangTask("OopStorageParIterPerf::Task"),
+ _worker_times(NULL),
+ _state(storage, nthreads),
+ _closure(closure)
+ {
+ Tickspan* wtimes = NEW_C_HEAP_ARRAY(Tickspan, _num_workers, mtInternal);
+ for (uint i = 0; i < _num_workers; ++i) {
+ new (&wtimes[i]) Tickspan();
+ }
+ _worker_times = wtimes;
+ }
+
+ ~Task() {
+ FREE_C_HEAP_ARRAY(Tickspan, _worker_times);
+ }
+
+ virtual void work(uint worker_id) {
+ Ticks start_time = Ticks::now();
+ _state.oops_do(_closure);
+ _worker_times[worker_id] = Ticks::now() - start_time;
+ }
+
+ const Tickspan* worker_times() const { return _worker_times; }
+};
+
+class OopStorageParIterPerf::Closure : public OopClosure {
+public:
+ virtual void do_oop(oop* p) { guarantee(*p == NULL, "expected NULL"); }
+ virtual void do_oop(narrowOop* p) { ShouldNotReachHere(); }
+};
+
+Tickspan OopStorageParIterPerf::run_task(Task* task, uint nthreads) {
+ tty->print_cr("Running test with %u threads", nthreads);
+ VM_ParStateTime op(workers(), task, nthreads);
+ ThreadInVMfromNative invm(JavaThread::current());
+ Ticks start_time = Ticks::now();
+ VMThread::execute(&op);
+ return Ticks::now() - start_time;
+}
+
+void OopStorageParIterPerf::show_task(const Task* task, Tickspan duration, uint nthreads) {
+ tty->print_cr("Run test with %u threads: " JLONG_FORMAT, nthreads, duration.value());
+ const Tickspan* wtimes = task->worker_times();
+ for (uint i = 0; i < _num_workers; ++i) {
+ if (wtimes[i] != Tickspan()) {
+ tty->print_cr(" %u: " JLONG_FORMAT, i, wtimes[i].value());
+ }
+ }
+ tty->cr();
+}
+
+void OopStorageParIterPerf::run_test(uint nthreads) {
+ if (nthreads <= _num_workers) {
+ SCOPED_TRACE(err_msg("Running test with %u threads", nthreads).buffer());
+ Closure closure;
+ Task task(&_storage, &closure, nthreads);
+ Tickspan t = run_task(&task, nthreads);
+ show_task(&task, t, nthreads);
+ }
+}
+
+TEST_VM_F(OopStorageParIterPerf, test) {
+ // Enable additional interesting logging.
+#define TEST_TAGS oopstorage, blocks, stats
+ // There isn't an obvious way to capture the old log level so it
+ // can be restored here, so just use Warning as the "default".
+ LogLevelType old_level = LogLevel::Warning;
+ if (log_is_enabled(Debug, TEST_TAGS)) {
+ old_level = LogLevel::Debug;
+ } else if (log_is_enabled(Info, TEST_TAGS)) {
+ old_level = LogLevel::Info;
+ }
+ bool debug_enabled = old_level == LogLevel::Debug;
+ if (!debug_enabled) {
+ LogConfiguration::configure_stdout(LogLevel::Debug, true, LOG_TAGS(TEST_TAGS));
+ }
+
+ run_test(1);
+ run_test(2);
+ run_test(3);
+ run_test(4);
+ run_test(6);
+ run_test(8);
+ run_test(10);
+
+ if (!debug_enabled) {
+ LogConfiguration::configure_stdout(old_level, true, LOG_TAGS(TEST_TAGS));
+ }
+}
+
+#endif // INCLUDE_ALL_GCS