stan/jdk-24
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

6692906: CMS: parallel concurrent marking may be prone to hanging or stalling mutators for periods of time

Browse Source 
Inserted missing yield(check)s in closures used during the work-stealing phase of parallel concurrent marking, a missing synchronous yield-request in the cms perm gen allocation path, and a terminator-terminator for the offer_termination invocation that monitors the yield status of the concurrent marking task. Elaborated some documentation comments and made some task queue termination loop flags configurable at start-up to aid debugging in the field.

Reviewed-by: jmasa, johnc, poonam
This commit is contained in:
Y. Srinivas Ramakrishna 2010-09-29 16:17:02 -07:00
parent 1cdd538ea5
commit cf2163334f
4 changed files with 97 additions and 34 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

68
hotspot/src/share/vm/gc_implementation/concurrentMarkSweep/concurrentMarkSweepGeneration.cpp
View File

@ -3264,6 +3264,7 @@ HeapWord*
ConcurrentMarkSweepGeneration::expand_and_allocate(size_t word_size, ConcurrentMarkSweepGeneration::expand_and_allocate(size_t word_size,
bool tlab, bool tlab,
bool parallel) { bool parallel) {
CMSSynchronousYieldRequest yr;
assert(!tlab, "Can't deal with TLAB allocation"); assert(!tlab, "Can't deal with TLAB allocation");
MutexLockerEx x(freelistLock(), Mutex::_no_safepoint_check_flag); MutexLockerEx x(freelistLock(), Mutex::_no_safepoint_check_flag);
expand(word_size*HeapWordSize, MinHeapDeltaBytes, expand(word_size*HeapWordSize, MinHeapDeltaBytes,
@ -3710,25 +3711,31 @@ class CMSConcMarkingTask;
class CMSConcMarkingTerminator: public ParallelTaskTerminator { class CMSConcMarkingTerminator: public ParallelTaskTerminator {
CMSCollector* _collector; CMSCollector* _collector;
CMSConcMarkingTask* _task; CMSConcMarkingTask* _task;
bool _yield;
protected:
virtual void yield();
public: public:
virtual void yield();
// "n_threads" is the number of threads to be terminated. // "n_threads" is the number of threads to be terminated.
// "queue_set" is a set of work queues of other threads. // "queue_set" is a set of work queues of other threads.
// "collector" is the CMS collector associated with this task terminator. // "collector" is the CMS collector associated with this task terminator.
// "yield" indicates whether we need the gang as a whole to yield. // "yield" indicates whether we need the gang as a whole to yield.
CMSConcMarkingTerminator(int n_threads, TaskQueueSetSuper* queue_set, CMSConcMarkingTerminator(int n_threads, TaskQueueSetSuper* queue_set, CMSCollector* collector) :
CMSCollector* collector, bool yield) :
ParallelTaskTerminator(n_threads, queue_set), ParallelTaskTerminator(n_threads, queue_set),
_collector(collector), _collector(collector) { }
_yield(yield) { }
void set_task(CMSConcMarkingTask* task) { void set_task(CMSConcMarkingTask* task) {
_task = task; _task = task;
} }
}; };
class CMSConcMarkingTerminatorTerminator: public TerminatorTerminator {
CMSConcMarkingTask* _task;
public:
bool should_exit_termination();
void set_task(CMSConcMarkingTask* task) {
_task = task;
}
};
// MT Concurrent Marking Task // MT Concurrent Marking Task
class CMSConcMarkingTask: public YieldingFlexibleGangTask { class CMSConcMarkingTask: public YieldingFlexibleGangTask {
CMSCollector* _collector; CMSCollector* _collector;
@ -3737,7 +3744,9 @@ class CMSConcMarkingTask: public YieldingFlexibleGangTask {
bool _result; bool _result;
CompactibleFreeListSpace* _cms_space; CompactibleFreeListSpace* _cms_space;
CompactibleFreeListSpace* _perm_space; CompactibleFreeListSpace* _perm_space;
HeapWord* _global_finger; char _pad_front[64]; // padding to ...
HeapWord* _global_finger; // ... avoid sharing cache line
char _pad_back[64];
HeapWord* _restart_addr; HeapWord* _restart_addr;
// Exposed here for yielding support // Exposed here for yielding support
@ -3745,7 +3754,10 @@ class CMSConcMarkingTask: public YieldingFlexibleGangTask {
// The per thread work queues, available here for stealing // The per thread work queues, available here for stealing
OopTaskQueueSet* _task_queues; OopTaskQueueSet* _task_queues;
// Termination (and yielding) support
CMSConcMarkingTerminator _term; CMSConcMarkingTerminator _term;
CMSConcMarkingTerminatorTerminator _term_term;
public: public:
CMSConcMarkingTask(CMSCollector* collector, CMSConcMarkingTask(CMSCollector* collector,
@ -3760,11 +3772,12 @@ class CMSConcMarkingTask: public YieldingFlexibleGangTask {
_perm_space(perm_space), _perm_space(perm_space),
_asynch(asynch), _n_workers(0), _result(true), _asynch(asynch), _n_workers(0), _result(true),
_task_queues(task_queues), _task_queues(task_queues),
_term(_n_workers, task_queues, _collector, asynch), _term(_n_workers, task_queues, _collector),
_bit_map_lock(collector->bitMapLock()) _bit_map_lock(collector->bitMapLock())
{ {
_requested_size = _n_workers; _requested_size = _n_workers;
_term.set_task(this); _term.set_task(this);
_term_term.set_task(this);
assert(_cms_space->bottom() < _perm_space->bottom(), assert(_cms_space->bottom() < _perm_space->bottom(),
"Finger incorrectly initialized below"); "Finger incorrectly initialized below");
_restart_addr = _global_finger = _cms_space->bottom(); _restart_addr = _global_finger = _cms_space->bottom();
@ -3784,6 +3797,11 @@ class CMSConcMarkingTask: public YieldingFlexibleGangTask {
} }
void work(int i); void work(int i);
bool should_yield() {
return ConcurrentMarkSweepThread::should_yield()
&& !_collector->foregroundGCIsActive()
&& _asynch;
}
virtual void coordinator_yield(); // stuff done by coordinator virtual void coordinator_yield(); // stuff done by coordinator
bool result() { return _result; } bool result() { return _result; }
@ -3805,10 +3823,17 @@ class CMSConcMarkingTask: public YieldingFlexibleGangTask {
void bump_global_finger(HeapWord* f); void bump_global_finger(HeapWord* f);
}; };
bool CMSConcMarkingTerminatorTerminator::should_exit_termination() {
assert(_task != NULL, "Error");
return _task->yielding();
// Note that we do not need the disjunct || _task->should_yield() above
// because we want terminating threads to yield only if the task
// is already in the midst of yielding, which happens only after at least one
// thread has yielded.
}
void CMSConcMarkingTerminator::yield() { void CMSConcMarkingTerminator::yield() {
if (ConcurrentMarkSweepThread::should_yield() && if (_task->should_yield()) {
!_collector->foregroundGCIsActive() &&
_yield) {
_task->yield(); _task->yield();
} else { } else {
ParallelTaskTerminator::yield(); ParallelTaskTerminator::yield();
@ -4033,6 +4058,7 @@ void CMSConcMarkingTask::do_scan_and_mark(int i, CompactibleFreeListSpace* sp) {
class Par_ConcMarkingClosure: public Par_KlassRememberingOopClosure { class Par_ConcMarkingClosure: public Par_KlassRememberingOopClosure {
private: private:
CMSConcMarkingTask* _task;
MemRegion _span; MemRegion _span;
CMSBitMap* _bit_map; CMSBitMap* _bit_map;
CMSMarkStack* _overflow_stack; CMSMarkStack* _overflow_stack;
@ -4040,11 +4066,12 @@ class Par_ConcMarkingClosure: public Par_KlassRememberingOopClosure {
protected: protected:
DO_OOP_WORK_DEFN DO_OOP_WORK_DEFN
public: public:
Par_ConcMarkingClosure(CMSCollector* collector, OopTaskQueue* work_queue, Par_ConcMarkingClosure(CMSCollector* collector, CMSConcMarkingTask* task, OopTaskQueue* work_queue,
CMSBitMap* bit_map, CMSMarkStack* overflow_stack, CMSBitMap* bit_map, CMSMarkStack* overflow_stack,
CMSMarkStack* revisit_stack): CMSMarkStack* revisit_stack):
Par_KlassRememberingOopClosure(collector, NULL, revisit_stack), Par_KlassRememberingOopClosure(collector, NULL, revisit_stack),
_span(_collector->_span), _task(task),
_span(collector->_span),
_work_queue(work_queue), _work_queue(work_queue),
_bit_map(bit_map), _bit_map(bit_map),
_overflow_stack(overflow_stack) _overflow_stack(overflow_stack)
@ -4053,6 +4080,11 @@ class Par_ConcMarkingClosure: public Par_KlassRememberingOopClosure {
virtual void do_oop(narrowOop* p); virtual void do_oop(narrowOop* p);
void trim_queue(size_t max); void trim_queue(size_t max);
void handle_stack_overflow(HeapWord* lost); void handle_stack_overflow(HeapWord* lost);
void do_yield_check() {
if (_task->should_yield()) {
_task->yield();
}
}
}; };
// Grey object scanning during work stealing phase -- // Grey object scanning during work stealing phase --
@ -4096,6 +4128,7 @@ void Par_ConcMarkingClosure::do_oop(oop obj) {
handle_stack_overflow(addr); handle_stack_overflow(addr);
} }
} // Else, some other thread got there first } // Else, some other thread got there first
do_yield_check();
} }
} }
@ -4111,6 +4144,7 @@ void Par_ConcMarkingClosure::trim_queue(size_t max) {
assert(_span.contains((HeapWord*)new_oop), "Not in span"); assert(_span.contains((HeapWord*)new_oop), "Not in span");
assert(new_oop->is_parsable(), "Should be parsable"); assert(new_oop->is_parsable(), "Should be parsable");
new_oop->oop_iterate(this); // do_oop() above new_oop->oop_iterate(this); // do_oop() above
do_yield_check();
} }
} }
} }
@ -4138,7 +4172,7 @@ void CMSConcMarkingTask::do_work_steal(int i) {
CMSMarkStack* ovflw = &(_collector->_markStack); CMSMarkStack* ovflw = &(_collector->_markStack);
CMSMarkStack* revisit = &(_collector->_revisitStack); CMSMarkStack* revisit = &(_collector->_revisitStack);
int* seed = _collector->hash_seed(i); int* seed = _collector->hash_seed(i);
Par_ConcMarkingClosure cl(_collector, work_q, bm, ovflw, revisit); Par_ConcMarkingClosure cl(_collector, this, work_q, bm, ovflw, revisit);
while (true) { while (true) {
cl.trim_queue(0); cl.trim_queue(0);
assert(work_q->size() == 0, "Should have been emptied above"); assert(work_q->size() == 0, "Should have been emptied above");
@ -4151,9 +4185,11 @@ void CMSConcMarkingTask::do_work_steal(int i) {
assert(obj_to_scan->is_oop(), "Should be an oop"); assert(obj_to_scan->is_oop(), "Should be an oop");
assert(bm->isMarked((HeapWord*)obj_to_scan), "Grey object"); assert(bm->isMarked((HeapWord*)obj_to_scan), "Grey object");
obj_to_scan->oop_iterate(&cl); obj_to_scan->oop_iterate(&cl);
} else if (terminator()->offer_termination()) { } else if (terminator()->offer_termination(&_term_term)) {
assert(work_q->size() == 0, "Impossible!"); assert(work_q->size() == 0, "Impossible!");
break; break;
} else if (yielding() || should_yield()) {
yield();
} }
} }
} }

48
hotspot/src/share/vm/gc_implementation/concurrentMarkSweep/concurrentMarkSweepThread.hpp
View File

@ -1,5 +1,5 @@
/* /*
* Copyright (c) 2001, 2006, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2001, 2010, Oracle and/or its affiliates. All rights reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
* *
* This code is free software; you can redistribute it and/or modify it * This code is free software; you can redistribute it and/or modify it
@ -123,24 +123,44 @@ class ConcurrentMarkSweepThread: public ConcurrentGCThread {
// or given timeout, whichever is earlier. // or given timeout, whichever is earlier.
void wait_on_cms_lock(long t); // milliseconds void wait_on_cms_lock(long t); // milliseconds
// The CMS thread will yield during the work portion of it's cycle // The CMS thread will yield during the work portion of its cycle
// only when requested to. Both synchronous and asychronous requests // only when requested to. Both synchronous and asychronous requests
// are provided. A synchronous request is used for young gen // are provided:
// collections and direct allocations. The requesting thread increments // (1) A synchronous request is used for young gen collections and
// pending_yields at the beginning of an operation, and decrements it when // for direct allocations. The requesting thread increments
// the operation is completed. The CMS thread yields when pending_yields // _pending_yields at the beginning of an operation, and decrements
// is positive. An asynchronous request is used by iCMS in the stop_icms() // _pending_yields when that operation is completed.
// operation. A single yield satisfies the outstanding asynch yield requests. // In turn, the CMS thread yields when _pending_yields is positive,
// The requesting thread increments both pending_yields and pending_decrements. // and continues to yield until the value reverts to 0.
// After yielding, the CMS thread decrements both by the amount in // (2) An asynchronous request, on the other hand, is used by iCMS
// pending_decrements. // for the stop_icms() operation. A single yield satisfies all of
// the outstanding asynch yield requests, of which there may
// occasionally be several in close succession. To accomplish
// this, an asynch-requesting thread atomically increments both
// _pending_yields and _pending_decrements. An asynchr requesting
// thread does not wait and "acknowledge" completion of an operation
// and deregister the request, like the synchronous version described
// above does. In turn, after yielding, the CMS thread decrements both
// _pending_yields and _pending_decrements by the value seen in
// _pending_decrements before the decrement.
// NOTE: The above scheme is isomorphic to having two request counters,
// one for async requests and one for sync requests, and for the CMS thread
// to check the sum of the two counters to decide whether it should yield
// and to clear only the async counter when it yields. However, it turns out
// to be more efficient for CMS code to just check a single counter
// _pending_yields that holds the sum (of both sync and async requests), and
// a second counter _pending_decrements that only holds the async requests,
// for greater efficiency, since in a typical CMS run, there are many more
// pontential (i.e. static) yield points than there are actual
// (i.e. dynamic) yields because of requests, which are few and far between.
//
// Note that, while "_pending_yields >= _pending_decrements" is an invariant, // Note that, while "_pending_yields >= _pending_decrements" is an invariant,
// we cannot easily test that invariant, since the counters are manipulated via // we cannot easily test that invariant, since the counters are manipulated via
// atomic instructions without explicit locking and we cannot read // atomic instructions without explicit locking and we cannot read
// the two counters atomically together: one suggestion is to // the two counters atomically together: one suggestion is to
// use (for example) 16-bit counters so as to be able to read the // use (for example) 16-bit counters so as to be able to read the
// two counters atomically even on 32-bit platforms. Notice that // two counters atomically even on 32-bit platforms. Notice that
// the second assert in acknowledge_yield_request() does indeed // the second assert in acknowledge_yield_request() below does indeed
// check a form of the above invariant, albeit indirectly. // check a form of the above invariant, albeit indirectly.
static void increment_pending_yields() { static void increment_pending_yields() {
@ -152,6 +172,7 @@ class ConcurrentMarkSweepThread: public ConcurrentGCThread {
assert(_pending_yields >= 0, "can't be negative"); assert(_pending_yields >= 0, "can't be negative");
} }
static void asynchronous_yield_request() { static void asynchronous_yield_request() {
assert(CMSIncrementalMode, "Currently only used w/iCMS");
increment_pending_yields(); increment_pending_yields();
Atomic::inc(&_pending_decrements); Atomic::inc(&_pending_decrements);
assert(_pending_decrements >= 0, "can't be negative"); assert(_pending_decrements >= 0, "can't be negative");
@ -159,6 +180,7 @@ class ConcurrentMarkSweepThread: public ConcurrentGCThread {
static void acknowledge_yield_request() { static void acknowledge_yield_request() {
jint decrement = _pending_decrements; jint decrement = _pending_decrements;
if (decrement > 0) { if (decrement > 0) {
assert(CMSIncrementalMode, "Currently only used w/iCMS");
// Order important to preserve: _pending_yields >= _pending_decrements // Order important to preserve: _pending_yields >= _pending_decrements
Atomic::add(-decrement, &_pending_decrements); Atomic::add(-decrement, &_pending_decrements);
Atomic::add(-decrement, &_pending_yields); Atomic::add(-decrement, &_pending_yields);
@ -195,7 +217,7 @@ inline void ConcurrentMarkSweepThread::trace_state(const char* desc) {
} }
} }
// For scoped increment/decrement of yield requests // For scoped increment/decrement of (synchronous) yield requests
class CMSSynchronousYieldRequest: public StackObj { class CMSSynchronousYieldRequest: public StackObj {
public: public:
CMSSynchronousYieldRequest() { CMSSynchronousYieldRequest() {

12
hotspot/src/share/vm/runtime/globals.hpp
View File

@ -280,6 +280,10 @@ class CommandLineFlags {
// UnlockExperimentalVMOptions flag, which allows the control and // UnlockExperimentalVMOptions flag, which allows the control and
// modification of the experimental flags. // modification of the experimental flags.
// //
// Nota bene: neither diagnostic nor experimental options should be used casually,
// and they are not supported on production loads, except under explicit
// direction from support engineers.
//
// manageable flags are writeable external product flags. // manageable flags are writeable external product flags.
// They are dynamically writeable through the JDK management interface // They are dynamically writeable through the JDK management interface
// (com.sun.management.HotSpotDiagnosticMXBean API) and also through JConsole. // (com.sun.management.HotSpotDiagnosticMXBean API) and also through JConsole.
@ -1809,17 +1813,17 @@ class CommandLineFlags {
develop(uintx, PromotionFailureALotInterval, 5, \ develop(uintx, PromotionFailureALotInterval, 5, \
"Total collections between promotion failures alot") \ "Total collections between promotion failures alot") \
\ \
develop(intx, WorkStealingSleepMillis, 1, \ experimental(intx, WorkStealingSleepMillis, 1, \
"Sleep time when sleep is used for yields") \ "Sleep time when sleep is used for yields") \
\ \
develop(uintx, WorkStealingYieldsBeforeSleep, 1000, \ experimental(uintx, WorkStealingYieldsBeforeSleep, 1000, \
"Number of yields before a sleep is done during workstealing") \ "Number of yields before a sleep is done during workstealing") \
\ \
develop(uintx, WorkStealingHardSpins, 4096, \ experimental(uintx, WorkStealingHardSpins, 4096, \
"Number of iterations in a spin loop between checks on " \ "Number of iterations in a spin loop between checks on " \
"time out of hard spin") \ "time out of hard spin") \
\ \
develop(uintx, WorkStealingSpinToYieldRatio, 10, \ experimental(uintx, WorkStealingSpinToYieldRatio, 10, \
"Ratio of hard spins to calls to yield") \ "Ratio of hard spins to calls to yield") \
\ \
product(uintx, PreserveMarkStackSize, 1024, \ product(uintx, PreserveMarkStackSize, 1024, \

3
hotspot/src/share/vm/utilities/yieldingWorkgroup.hpp
View File

@ -1,5 +1,5 @@
/* /*
* Copyright (c) 2005, 2010 Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
* *
* This code is free software; you can redistribute it and/or modify it * This code is free software; you can redistribute it and/or modify it
@ -137,6 +137,7 @@ public:
virtual void abort(); virtual void abort();
Status status() const { return _status; } Status status() const { return _status; }
bool yielding() const { return _status == YIELDING; }
bool yielded() const { return _status == YIELDED; } bool yielded() const { return _status == YIELDED; }
bool completed() const { return _status == COMPLETED; } bool completed() const { return _status == COMPLETED; }
bool aborted() const { return _status == ABORTED; } bool aborted() const { return _status == ABORTED; }