Mercurial > hg > truffle
view src/share/vm/utilities/events.cpp @ 1994:6cd6d394f280
7001033: assert(gch->gc_cause() == GCCause::_scavenge_alot || !gch->incremental_collection_failed())
7002546: regression on SpecJbb2005 on 7b118 comparing to 7b117 on small heaps
Summary: Relaxed assertion checking related to incremental_collection_failed flag to allow for ExplicitGCInvokesConcurrent behaviour where we do not want a failing scavenge to bail to a stop-world collection. Parameterized incremental_collection_will_fail() so we can selectively use, or not use, as appropriate, the statistical prediction at specific use sites. This essentially reverts the scavenge bail-out logic to what it was prior to some recent changes that had inadvertently started using the statistical prediction which can be noisy in the presence of bursty loads. Added some associated verbose non-product debugging messages.
Reviewed-by: johnc, tonyp
author | ysr |
---|---|
date | Tue, 07 Dec 2010 21:55:53 -0800 |
parents | f95d63e2154a |
children | f08d439fab8c |
line wrap: on
line source
/* * Copyright (c) 1997, 2010, 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 "memory/allocation.inline.hpp" #include "runtime/mutexLocker.hpp" #include "runtime/osThread.hpp" #include "runtime/threadLocalStorage.hpp" #include "runtime/timer.hpp" #include "utilities/events.hpp" #ifdef TARGET_OS_FAMILY_linux # include "thread_linux.inline.hpp" #endif #ifdef TARGET_OS_FAMILY_solaris # include "thread_solaris.inline.hpp" #endif #ifdef TARGET_OS_FAMILY_windows # include "thread_windows.inline.hpp" #endif #ifndef PRODUCT //////////////////////////////////////////////////////////////////////////// // Event typedef u4 EventID; class Event VALUE_OBJ_CLASS_SPEC { private: jlong _time_tick; intx _thread_id; const char* _format; int _indent; intptr_t _arg_1; intptr_t _arg_2; intptr_t _arg_3; // only EventBuffer::add_event() can assign event id friend class EventBuffer; EventID _id; public: void clear() { _format = NULL; } EventID id() const { return _id; } void fill(int indent, const char* format, intptr_t arg_1, intptr_t arg_2, intptr_t arg_3) { _format = format; _arg_1 = arg_1; _arg_2 = arg_2; _arg_3 = arg_3; _indent = indent; _thread_id = os::current_thread_id(); _time_tick = os::elapsed_counter(); } void print_on(outputStream *st) { if (_format == NULL) return; st->print(" %d", _thread_id); st->print(" %3.2g ", (double)_time_tick / os::elapsed_frequency()); st->fill_to(20); for (int index = 0; index < _indent; index++) { st->print("| "); } st->print_cr(_format, _arg_1, _arg_2, _arg_3); } }; //////////////////////////////////////////////////////////////////////////// // EventBuffer // // Simple lock-free event queue. Every event has a unique 32-bit id. // It's fine if two threads add events at the same time, because they // will get different event id, and then write to different buffer location. // However, it is assumed that add_event() is quick enough (or buffer size // is big enough), so when one thread is adding event, there can't be more // than "size" events created by other threads; otherwise we'll end up having // two threads writing to the same location. class EventBuffer : AllStatic { private: static Event* buffer; static int size; static jint indent; static volatile EventID _current_event_id; static EventID get_next_event_id() { return (EventID)Atomic::add(1, (jint*)&_current_event_id); } public: static void inc_indent() { Atomic::inc(&indent); } static void dec_indent() { Atomic::dec(&indent); } static bool get_event(EventID id, Event* event) { int index = (int)(id % size); if (buffer[index].id() == id) { memcpy(event, &buffer[index], sizeof(Event)); // check id again; if buffer[index] is being updated by another thread, // event->id() will contain different value. return (event->id() == id); } else { // id does not match - id is invalid, or event is overwritten return false; } } // add a new event to the queue; if EventBuffer is full, this call will // overwrite the oldest event in the queue static EventID add_event(const char* format, intptr_t arg_1, intptr_t arg_2, intptr_t arg_3) { // assign a unique id EventID id = get_next_event_id(); // event will be copied to buffer[index] int index = (int)(id % size); // first, invalidate id, buffer[index] can't have event with id = index + 2 buffer[index]._id = index + 2; // make sure everyone has seen that buffer[index] is invalid OrderAccess::fence(); // ... before updating its value buffer[index].fill(indent, format, arg_1, arg_2, arg_3); // finally, set up real event id, now buffer[index] contains valid event OrderAccess::release_store(&(buffer[index]._id), id); return id; } static void print_last(outputStream *st, int number) { st->print_cr("[Last %d events in the event buffer]", number); st->print_cr("-<thd>-<elapsed sec>-<description>---------------------"); int count = 0; EventID id = _current_event_id; while (count < number) { Event event; if (get_event(id, &event)) { event.print_on(st); } id--; count++; } } static void print_all(outputStream* st) { print_last(st, size); } static void init() { // Allocate the event buffer size = EventLogLength; buffer = NEW_C_HEAP_ARRAY(Event, size); _current_event_id = 0; // Clear the event buffer for (int index = 0; index < size; index++) { buffer[index]._id = index + 1; // index + 1 is invalid id buffer[index].clear(); } } }; Event* EventBuffer::buffer; int EventBuffer::size; volatile EventID EventBuffer::_current_event_id; int EventBuffer::indent; //////////////////////////////////////////////////////////////////////////// // Events // Events::log() is safe for signal handlers void Events::log(const char* format, ...) { if (LogEvents) { va_list ap; va_start(ap, format); intptr_t arg_1 = va_arg(ap, intptr_t); intptr_t arg_2 = va_arg(ap, intptr_t); intptr_t arg_3 = va_arg(ap, intptr_t); va_end(ap); EventBuffer::add_event(format, arg_1, arg_2, arg_3); } } void Events::print_all(outputStream *st) { EventBuffer::print_all(st); } void Events::print_last(outputStream *st, int number) { EventBuffer::print_last(st, number); } /////////////////////////////////////////////////////////////////////////// // EventMark EventMark::EventMark(const char* format, ...) { if (LogEvents) { va_list ap; va_start(ap, format); intptr_t arg_1 = va_arg(ap, intptr_t); intptr_t arg_2 = va_arg(ap, intptr_t); intptr_t arg_3 = va_arg(ap, intptr_t); va_end(ap); EventBuffer::add_event(format, arg_1, arg_2, arg_3); EventBuffer::inc_indent(); } } EventMark::~EventMark() { if (LogEvents) { EventBuffer::dec_indent(); EventBuffer::add_event("done", 0, 0, 0); } } /////////////////////////////////////////////////////////////////////////// void eventlog_init() { EventBuffer::init(); } int print_all_events(outputStream *st) { EventBuffer::print_all(st); return 1; } #else void eventlog_init() {} int print_all_events(outputStream *st) { return 0; } #endif // PRODUCT