3 * Simple generational GC.
6 * Paolo Molaro (lupus@ximian.com)
7 * Rodrigo Kumpera (kumpera@gmail.com)
9 * Copyright 2005-2011 Novell, Inc (http://www.novell.com)
10 * Copyright 2011 Xamarin Inc (http://www.xamarin.com)
12 * Thread start/stop adapted from Boehm's GC:
13 * Copyright (c) 1994 by Xerox Corporation. All rights reserved.
14 * Copyright (c) 1996 by Silicon Graphics. All rights reserved.
15 * Copyright (c) 1998 by Fergus Henderson. All rights reserved.
16 * Copyright (c) 2000-2004 by Hewlett-Packard Company. All rights reserved.
17 * Copyright 2001-2003 Ximian, Inc
18 * Copyright 2003-2010 Novell, Inc.
19 * Copyright 2011 Xamarin, Inc.
20 * Copyright (C) 2012 Xamarin Inc
22 * Licensed under the MIT license. See LICENSE file in the project root for full license information.
24 * Important: allocation provides always zeroed memory, having to do
25 * a memset after allocation is deadly for performance.
26 * Memory usage at startup is currently as follows:
28 * 64 KB internal space
30 * We should provide a small memory config with half the sizes
32 * We currently try to make as few mono assumptions as possible:
33 * 1) 2-word header with no GC pointers in it (first vtable, second to store the
35 * 2) gc descriptor is the second word in the vtable (first word in the class)
36 * 3) 8 byte alignment is the minimum and enough (not true for special structures (SIMD), FIXME)
37 * 4) there is a function to get an object's size and the number of
38 * elements in an array.
39 * 5) we know the special way bounds are allocated for complex arrays
40 * 6) we know about proxies and how to treat them when domains are unloaded
42 * Always try to keep stack usage to a minimum: no recursive behaviour
43 * and no large stack allocs.
45 * General description.
46 * Objects are initially allocated in a nursery using a fast bump-pointer technique.
47 * When the nursery is full we start a nursery collection: this is performed with a
49 * When the old generation is full we start a copying GC of the old generation as well:
50 * this will be changed to mark&sweep with copying when fragmentation becomes to severe
51 * in the future. Maybe we'll even do both during the same collection like IMMIX.
53 * The things that complicate this description are:
54 * *) pinned objects: we can't move them so we need to keep track of them
55 * *) no precise info of the thread stacks and registers: we need to be able to
56 * quickly find the objects that may be referenced conservatively and pin them
57 * (this makes the first issues more important)
58 * *) large objects are too expensive to be dealt with using copying GC: we handle them
59 * with mark/sweep during major collections
60 * *) some objects need to not move even if they are small (interned strings, Type handles):
61 * we use mark/sweep for them, too: they are not allocated in the nursery, but inside
62 * PinnedChunks regions
68 *) we could have a function pointer in MonoClass to implement
69 customized write barriers for value types
71 *) investigate the stuff needed to advance a thread to a GC-safe
72 point (single-stepping, read from unmapped memory etc) and implement it.
73 This would enable us to inline allocations and write barriers, for example,
74 or at least parts of them, like the write barrier checks.
75 We may need this also for handling precise info on stacks, even simple things
76 as having uninitialized data on the stack and having to wait for the prolog
77 to zero it. Not an issue for the last frame that we scan conservatively.
78 We could always not trust the value in the slots anyway.
80 *) modify the jit to save info about references in stack locations:
81 this can be done just for locals as a start, so that at least
82 part of the stack is handled precisely.
84 *) test/fix endianess issues
86 *) Implement a card table as the write barrier instead of remembered
87 sets? Card tables are not easy to implement with our current
88 memory layout. We have several different kinds of major heap
89 objects: Small objects in regular blocks, small objects in pinned
90 chunks and LOS objects. If we just have a pointer we have no way
91 to tell which kind of object it points into, therefore we cannot
92 know where its card table is. The least we have to do to make
93 this happen is to get rid of write barriers for indirect stores.
96 *) Get rid of write barriers for indirect stores. We can do this by
97 telling the GC to wbarrier-register an object once we do an ldloca
98 or ldelema on it, and to unregister it once it's not used anymore
99 (it can only travel downwards on the stack). The problem with
100 unregistering is that it needs to happen eventually no matter
101 what, even if exceptions are thrown, the thread aborts, etc.
102 Rodrigo suggested that we could do only the registering part and
103 let the collector find out (pessimistically) when it's safe to
104 unregister, namely when the stack pointer of the thread that
105 registered the object is higher than it was when the registering
106 happened. This might make for a good first implementation to get
107 some data on performance.
109 *) Some sort of blacklist support? Blacklists is a concept from the
110 Boehm GC: if during a conservative scan we find pointers to an
111 area which we might use as heap, we mark that area as unusable, so
112 pointer retention by random pinning pointers is reduced.
114 *) experiment with max small object size (very small right now - 2kb,
115 because it's tied to the max freelist size)
117 *) add an option to mmap the whole heap in one chunk: it makes for many
118 simplifications in the checks (put the nursery at the top and just use a single
119 check for inclusion/exclusion): the issue this has is that on 32 bit systems it's
120 not flexible (too much of the address space may be used by default or we can't
121 increase the heap as needed) and we'd need a race-free mechanism to return memory
122 back to the system (mprotect(PROT_NONE) will still keep the memory allocated if it
123 was written to, munmap is needed, but the following mmap may not find the same segment
126 *) memzero the major fragments after restarting the world and optionally a smaller
129 *) investigate having fragment zeroing threads
131 *) separate locks for finalization and other minor stuff to reduce
134 *) try a different copying order to improve memory locality
136 *) a thread abort after a store but before the write barrier will
137 prevent the write barrier from executing
139 *) specialized dynamically generated markers/copiers
141 *) Dynamically adjust TLAB size to the number of threads. If we have
142 too many threads that do allocation, we might need smaller TLABs,
143 and we might get better performance with larger TLABs if we only
144 have a handful of threads. We could sum up the space left in all
145 assigned TLABs and if that's more than some percentage of the
146 nursery size, reduce the TLAB size.
148 *) Explore placing unreachable objects on unused nursery memory.
149 Instead of memset'ng a region to zero, place an int[] covering it.
150 A good place to start is add_nursery_frag. The tricky thing here is
151 placing those objects atomically outside of a collection.
153 *) Allocation should use asymmetric Dekker synchronization:
154 http://blogs.oracle.com/dave/resource/Asymmetric-Dekker-Synchronization.txt
155 This should help weak consistency archs.
162 #define _XOPEN_SOURCE
163 #define _DARWIN_C_SOURCE
169 #ifdef HAVE_PTHREAD_H
172 #ifdef HAVE_PTHREAD_NP_H
173 #include <pthread_np.h>
181 #include "mono/sgen/sgen-gc.h"
182 #include "mono/sgen/sgen-cardtable.h"
183 #include "mono/sgen/sgen-protocol.h"
184 #include "mono/sgen/sgen-memory-governor.h"
185 #include "mono/sgen/sgen-hash-table.h"
186 #include "mono/sgen/sgen-pinning.h"
187 #include "mono/sgen/sgen-workers.h"
188 #include "mono/sgen/sgen-client.h"
189 #include "mono/sgen/sgen-pointer-queue.h"
190 #include "mono/sgen/gc-internal-agnostic.h"
191 #include "mono/utils/mono-proclib.h"
192 #include "mono/utils/mono-memory-model.h"
193 #include "mono/utils/hazard-pointer.h"
195 #include <mono/utils/memcheck.h>
196 #include <mono/utils/mono-mmap-internals.h>
198 #undef pthread_create
200 #undef pthread_detach
203 * ######################################################################
204 * ######## Types and constants used by the GC.
205 * ######################################################################
208 /* 0 means not initialized, 1 is initialized, -1 means in progress */
209 static int gc_initialized = 0;
210 /* If set, check if we need to do something every X allocations */
211 gboolean has_per_allocation_action;
212 /* If set, do a heap check every X allocation */
213 guint32 verify_before_allocs = 0;
214 /* If set, do a minor collection before every X allocation */
215 guint32 collect_before_allocs = 0;
216 /* If set, do a whole heap check before each collection */
217 static gboolean whole_heap_check_before_collection = FALSE;
218 /* If set, do a remset consistency check at various opportunities */
219 static gboolean remset_consistency_checks = FALSE;
220 /* If set, do a mod union consistency check before each finishing collection pause */
221 static gboolean mod_union_consistency_check = FALSE;
222 /* If set, check whether mark bits are consistent after major collections */
223 static gboolean check_mark_bits_after_major_collection = FALSE;
224 /* If set, check that all nursery objects are pinned/not pinned, depending on context */
225 static gboolean check_nursery_objects_pinned = FALSE;
226 /* If set, do a few checks when the concurrent collector is used */
227 static gboolean do_concurrent_checks = FALSE;
228 /* If set, do a plausibility check on the scan_starts before and after
230 static gboolean do_scan_starts_check = FALSE;
232 static gboolean disable_minor_collections = FALSE;
233 static gboolean disable_major_collections = FALSE;
234 static gboolean do_verify_nursery = FALSE;
235 static gboolean do_dump_nursery_content = FALSE;
236 static gboolean enable_nursery_canaries = FALSE;
238 static gboolean precleaning_enabled = TRUE;
240 #ifdef HEAVY_STATISTICS
241 guint64 stat_objects_alloced_degraded = 0;
242 guint64 stat_bytes_alloced_degraded = 0;
244 guint64 stat_copy_object_called_nursery = 0;
245 guint64 stat_objects_copied_nursery = 0;
246 guint64 stat_copy_object_called_major = 0;
247 guint64 stat_objects_copied_major = 0;
249 guint64 stat_scan_object_called_nursery = 0;
250 guint64 stat_scan_object_called_major = 0;
252 guint64 stat_slots_allocated_in_vain;
254 guint64 stat_nursery_copy_object_failed_from_space = 0;
255 guint64 stat_nursery_copy_object_failed_forwarded = 0;
256 guint64 stat_nursery_copy_object_failed_pinned = 0;
257 guint64 stat_nursery_copy_object_failed_to_space = 0;
259 static guint64 stat_wbarrier_add_to_global_remset = 0;
260 static guint64 stat_wbarrier_arrayref_copy = 0;
261 static guint64 stat_wbarrier_generic_store = 0;
262 static guint64 stat_wbarrier_generic_store_atomic = 0;
263 static guint64 stat_wbarrier_set_root = 0;
266 static guint64 stat_pinned_objects = 0;
268 static guint64 time_minor_pre_collection_fragment_clear = 0;
269 static guint64 time_minor_pinning = 0;
270 static guint64 time_minor_scan_remsets = 0;
271 static guint64 time_minor_scan_major_blocks = 0;
272 static guint64 time_minor_scan_los = 0;
273 static guint64 time_minor_scan_pinned = 0;
274 static guint64 time_minor_scan_roots = 0;
275 static guint64 time_minor_finish_gray_stack = 0;
276 static guint64 time_minor_fragment_creation = 0;
278 static guint64 time_major_pre_collection_fragment_clear = 0;
279 static guint64 time_major_pinning = 0;
280 static guint64 time_major_scan_pinned = 0;
281 static guint64 time_major_scan_roots = 0;
282 static guint64 time_major_scan_mod_union = 0;
283 static guint64 time_major_finish_gray_stack = 0;
284 static guint64 time_major_free_bigobjs = 0;
285 static guint64 time_major_los_sweep = 0;
286 static guint64 time_major_sweep = 0;
287 static guint64 time_major_fragment_creation = 0;
289 static guint64 time_max = 0;
291 static SGEN_TV_DECLARE (time_major_conc_collection_start);
292 static SGEN_TV_DECLARE (time_major_conc_collection_end);
294 int gc_debug_level = 0;
296 static char* gc_params_options;
297 static char* gc_debug_options;
301 mono_gc_flush_info (void)
303 fflush (gc_debug_file);
307 #define TV_DECLARE SGEN_TV_DECLARE
308 #define TV_GETTIME SGEN_TV_GETTIME
309 #define TV_ELAPSED SGEN_TV_ELAPSED
311 static SGEN_TV_DECLARE (sgen_init_timestamp);
313 NurseryClearPolicy nursery_clear_policy = CLEAR_AT_TLAB_CREATION;
315 #define object_is_forwarded SGEN_OBJECT_IS_FORWARDED
316 #define object_is_pinned SGEN_OBJECT_IS_PINNED
317 #define pin_object SGEN_PIN_OBJECT
319 #define ptr_in_nursery sgen_ptr_in_nursery
321 #define LOAD_VTABLE SGEN_LOAD_VTABLE
324 nursery_canaries_enabled (void)
326 return enable_nursery_canaries;
329 #define safe_object_get_size sgen_safe_object_get_size
331 #if defined(HAVE_CONC_GC_AS_DEFAULT)
332 /* Use concurrent major on deskstop platforms */
333 #define DEFAULT_MAJOR_INIT sgen_marksweep_conc_init
334 #define DEFAULT_MAJOR_NAME "marksweep-conc"
336 #define DEFAULT_MAJOR_INIT sgen_marksweep_init
337 #define DEFAULT_MAJOR_NAME "marksweep"
341 * ######################################################################
342 * ######## Global data.
343 * ######################################################################
345 MonoCoopMutex gc_mutex;
347 #define SCAN_START_SIZE SGEN_SCAN_START_SIZE
349 size_t degraded_mode = 0;
351 static mword bytes_pinned_from_failed_allocation = 0;
353 GCMemSection *nursery_section = NULL;
354 static volatile mword lowest_heap_address = ~(mword)0;
355 static volatile mword highest_heap_address = 0;
357 MonoCoopMutex sgen_interruption_mutex;
359 int current_collection_generation = -1;
360 static volatile gboolean concurrent_collection_in_progress = FALSE;
362 /* objects that are ready to be finalized */
363 static SgenPointerQueue fin_ready_queue = SGEN_POINTER_QUEUE_INIT (INTERNAL_MEM_FINALIZE_READY);
364 static SgenPointerQueue critical_fin_queue = SGEN_POINTER_QUEUE_INIT (INTERNAL_MEM_FINALIZE_READY);
366 /* registered roots: the key to the hash is the root start address */
368 * Different kinds of roots are kept separate to speed up pin_from_roots () for example.
370 SgenHashTable roots_hash [ROOT_TYPE_NUM] = {
371 SGEN_HASH_TABLE_INIT (INTERNAL_MEM_ROOTS_TABLE, INTERNAL_MEM_ROOT_RECORD, sizeof (RootRecord), sgen_aligned_addr_hash, NULL),
372 SGEN_HASH_TABLE_INIT (INTERNAL_MEM_ROOTS_TABLE, INTERNAL_MEM_ROOT_RECORD, sizeof (RootRecord), sgen_aligned_addr_hash, NULL),
373 SGEN_HASH_TABLE_INIT (INTERNAL_MEM_ROOTS_TABLE, INTERNAL_MEM_ROOT_RECORD, sizeof (RootRecord), sgen_aligned_addr_hash, NULL)
375 static mword roots_size = 0; /* amount of memory in the root set */
377 /* The size of a TLAB */
378 /* The bigger the value, the less often we have to go to the slow path to allocate a new
379 * one, but the more space is wasted by threads not allocating much memory.
381 * FIXME: Make this self-tuning for each thread.
383 guint32 tlab_size = (1024 * 4);
385 #define MAX_SMALL_OBJ_SIZE SGEN_MAX_SMALL_OBJ_SIZE
387 #define ALLOC_ALIGN SGEN_ALLOC_ALIGN
389 #define ALIGN_UP SGEN_ALIGN_UP
391 #ifdef SGEN_DEBUG_INTERNAL_ALLOC
392 MonoNativeThreadId main_gc_thread = NULL;
395 /*Object was pinned during the current collection*/
396 static mword objects_pinned;
399 * ######################################################################
400 * ######## Macros and function declarations.
401 * ######################################################################
404 /* forward declarations */
405 static void scan_from_registered_roots (char *addr_start, char *addr_end, int root_type, ScanCopyContext ctx);
407 static void pin_from_roots (void *start_nursery, void *end_nursery, ScanCopyContext ctx);
408 static void finish_gray_stack (int generation, ScanCopyContext ctx);
411 SgenMajorCollector major_collector;
412 SgenMinorCollector sgen_minor_collector;
414 static SgenRememberedSet remset;
417 * The gray queue a worker job must use. If we're not parallel or
418 * concurrent, we use the main gray queue.
420 static SgenGrayQueue*
421 sgen_workers_get_job_gray_queue (WorkerData *worker_data, SgenGrayQueue *default_gray_queue)
424 return &worker_data->private_gray_queue;
425 SGEN_ASSERT (0, default_gray_queue, "Why don't we have a default gray queue when we're not running in a worker thread?");
426 return default_gray_queue;
430 gray_queue_redirect (SgenGrayQueue *queue)
432 sgen_workers_take_from_queue (queue);
436 sgen_scan_area_with_callback (char *start, char *end, IterateObjectCallbackFunc callback, void *data, gboolean allow_flags, gboolean fail_on_canaries)
438 while (start < end) {
442 if (!*(void**)start) {
443 start += sizeof (void*); /* should be ALLOC_ALIGN, really */
448 if (!(obj = (char *)SGEN_OBJECT_IS_FORWARDED (start)))
454 if (!sgen_client_object_is_array_fill ((GCObject*)obj)) {
455 CHECK_CANARY_FOR_OBJECT ((GCObject*)obj, fail_on_canaries);
456 size = ALIGN_UP (safe_object_get_size ((GCObject*)obj));
457 callback ((GCObject*)obj, size, data);
458 CANARIFY_SIZE (size);
460 size = ALIGN_UP (safe_object_get_size ((GCObject*)obj));
468 * sgen_add_to_global_remset:
470 * The global remset contains locations which point into newspace after
471 * a minor collection. This can happen if the objects they point to are pinned.
473 * LOCKING: If called from a parallel collector, the global remset
474 * lock must be held. For serial collectors that is not necessary.
477 sgen_add_to_global_remset (gpointer ptr, GCObject *obj)
479 SGEN_ASSERT (5, sgen_ptr_in_nursery (obj), "Target pointer of global remset must be in the nursery");
481 HEAVY_STAT (++stat_wbarrier_add_to_global_remset);
483 if (!major_collector.is_concurrent) {
484 SGEN_ASSERT (5, current_collection_generation != -1, "Global remsets can only be added during collections");
486 if (current_collection_generation == -1)
487 SGEN_ASSERT (5, sgen_concurrent_collection_in_progress (), "Global remsets outside of collection pauses can only be added by the concurrent collector");
490 if (!object_is_pinned (obj))
491 SGEN_ASSERT (5, sgen_minor_collector.is_split || sgen_concurrent_collection_in_progress (), "Non-pinned objects can only remain in nursery if it is a split nursery");
492 else if (sgen_cement_lookup_or_register (obj))
495 remset.record_pointer (ptr);
497 sgen_pin_stats_register_global_remset (obj);
499 SGEN_LOG (8, "Adding global remset for %p", ptr);
500 binary_protocol_global_remset (ptr, obj, (gpointer)SGEN_LOAD_VTABLE (obj));
504 * sgen_drain_gray_stack:
506 * Scan objects in the gray stack until the stack is empty. This should be called
507 * frequently after each object is copied, to achieve better locality and cache
512 sgen_drain_gray_stack (ScanCopyContext ctx)
514 SGEN_ASSERT (0, ctx.ops->drain_gray_stack, "Why do we have a scan/copy context with a missing drain gray stack function?");
516 return ctx.ops->drain_gray_stack (ctx.queue);
520 * Addresses in the pin queue are already sorted. This function finds
521 * the object header for each address and pins the object. The
522 * addresses must be inside the nursery section. The (start of the)
523 * address array is overwritten with the addresses of the actually
524 * pinned objects. Return the number of pinned objects.
527 pin_objects_from_nursery_pin_queue (gboolean do_scan_objects, ScanCopyContext ctx)
529 GCMemSection *section = nursery_section;
530 void **start = sgen_pinning_get_entry (section->pin_queue_first_entry);
531 void **end = sgen_pinning_get_entry (section->pin_queue_last_entry);
532 void *start_nursery = section->data;
533 void *end_nursery = section->next_data;
538 void *pinning_front = start_nursery;
540 void **definitely_pinned = start;
541 ScanObjectFunc scan_func = ctx.ops->scan_object;
542 SgenGrayQueue *queue = ctx.queue;
544 sgen_nursery_allocator_prepare_for_pinning ();
546 while (start < end) {
547 GCObject *obj_to_pin = NULL;
548 size_t obj_to_pin_size = 0;
553 SGEN_ASSERT (0, addr >= start_nursery && addr < end_nursery, "Potential pinning address out of range");
554 SGEN_ASSERT (0, addr >= last, "Pin queue not sorted");
561 SGEN_LOG (5, "Considering pinning addr %p", addr);
562 /* We've already processed everything up to pinning_front. */
563 if (addr < pinning_front) {
569 * Find the closest scan start <= addr. We might search backward in the
570 * scan_starts array because entries might be NULL. In the worst case we
571 * start at start_nursery.
573 idx = ((char*)addr - (char*)section->data) / SCAN_START_SIZE;
574 SGEN_ASSERT (0, idx < section->num_scan_start, "Scan start index out of range");
575 search_start = (void*)section->scan_starts [idx];
576 if (!search_start || search_start > addr) {
579 search_start = section->scan_starts [idx];
580 if (search_start && search_start <= addr)
583 if (!search_start || search_start > addr)
584 search_start = start_nursery;
588 * If the pinning front is closer than the scan start we found, start
589 * searching at the front.
591 if (search_start < pinning_front)
592 search_start = pinning_front;
595 * Now addr should be in an object a short distance from search_start.
597 * search_start must point to zeroed mem or point to an object.
600 size_t obj_size, canarified_obj_size;
603 if (!*(void**)search_start) {
604 search_start = (void*)ALIGN_UP ((mword)search_start + sizeof (gpointer));
605 /* The loop condition makes sure we don't overrun addr. */
609 canarified_obj_size = obj_size = ALIGN_UP (safe_object_get_size ((GCObject*)search_start));
612 * Filler arrays are marked by an invalid sync word. We don't
613 * consider them for pinning. They are not delimited by canaries,
616 if (!sgen_client_object_is_array_fill ((GCObject*)search_start)) {
617 CHECK_CANARY_FOR_OBJECT (search_start, TRUE);
618 CANARIFY_SIZE (canarified_obj_size);
620 if (addr >= search_start && (char*)addr < (char*)search_start + obj_size) {
621 /* This is the object we're looking for. */
622 obj_to_pin = (GCObject*)search_start;
623 obj_to_pin_size = canarified_obj_size;
628 /* Skip to the next object */
629 search_start = (void*)((char*)search_start + canarified_obj_size);
630 } while (search_start <= addr);
632 /* We've searched past the address we were looking for. */
634 pinning_front = search_start;
635 goto next_pin_queue_entry;
639 * We've found an object to pin. It might still be a dummy array, but we
640 * can advance the pinning front in any case.
642 pinning_front = (char*)obj_to_pin + obj_to_pin_size;
645 * If this is a dummy array marking the beginning of a nursery
646 * fragment, we don't pin it.
648 if (sgen_client_object_is_array_fill (obj_to_pin))
649 goto next_pin_queue_entry;
652 * Finally - pin the object!
654 desc = sgen_obj_get_descriptor_safe (obj_to_pin);
655 if (do_scan_objects) {
656 scan_func (obj_to_pin, desc, queue);
658 SGEN_LOG (4, "Pinned object %p, vtable %p (%s), count %d\n",
659 obj_to_pin, *(void**)obj_to_pin, sgen_client_vtable_get_name (SGEN_LOAD_VTABLE (obj_to_pin)), count);
660 binary_protocol_pin (obj_to_pin,
661 (gpointer)LOAD_VTABLE (obj_to_pin),
662 safe_object_get_size (obj_to_pin));
664 pin_object (obj_to_pin);
665 GRAY_OBJECT_ENQUEUE_SERIAL (queue, obj_to_pin, desc);
666 sgen_pin_stats_register_object (obj_to_pin, GENERATION_NURSERY);
667 definitely_pinned [count] = obj_to_pin;
670 if (concurrent_collection_in_progress)
671 sgen_pinning_register_pinned_in_nursery (obj_to_pin);
673 next_pin_queue_entry:
677 sgen_client_nursery_objects_pinned (definitely_pinned, count);
678 stat_pinned_objects += count;
683 pin_objects_in_nursery (gboolean do_scan_objects, ScanCopyContext ctx)
687 if (nursery_section->pin_queue_first_entry == nursery_section->pin_queue_last_entry)
690 reduced_to = pin_objects_from_nursery_pin_queue (do_scan_objects, ctx);
691 nursery_section->pin_queue_last_entry = nursery_section->pin_queue_first_entry + reduced_to;
695 * This function is only ever called (via `collector_pin_object()` in `sgen-copy-object.h`)
696 * when we can't promote an object because we're out of memory.
699 sgen_pin_object (GCObject *object, SgenGrayQueue *queue)
701 SGEN_ASSERT (0, sgen_ptr_in_nursery (object), "We're only supposed to use this for pinning nursery objects when out of memory.");
704 * All pinned objects are assumed to have been staged, so we need to stage as well.
705 * Also, the count of staged objects shows that "late pinning" happened.
707 sgen_pin_stage_ptr (object);
709 SGEN_PIN_OBJECT (object);
710 binary_protocol_pin (object, (gpointer)LOAD_VTABLE (object), safe_object_get_size (object));
713 sgen_pin_stats_register_object (object, GENERATION_NURSERY);
715 GRAY_OBJECT_ENQUEUE_SERIAL (queue, object, sgen_obj_get_descriptor_safe (object));
718 /* Sort the addresses in array in increasing order.
719 * Done using a by-the book heap sort. Which has decent and stable performance, is pretty cache efficient.
722 sgen_sort_addresses (void **array, size_t size)
727 for (i = 1; i < size; ++i) {
730 size_t parent = (child - 1) / 2;
732 if (array [parent] >= array [child])
735 tmp = array [parent];
736 array [parent] = array [child];
743 for (i = size - 1; i > 0; --i) {
746 array [i] = array [0];
752 while (root * 2 + 1 <= end) {
753 size_t child = root * 2 + 1;
755 if (child < end && array [child] < array [child + 1])
757 if (array [root] >= array [child])
761 array [root] = array [child];
770 * Scan the memory between start and end and queue values which could be pointers
771 * to the area between start_nursery and end_nursery for later consideration.
772 * Typically used for thread stacks.
775 sgen_conservatively_pin_objects_from (void **start, void **end, void *start_nursery, void *end_nursery, int pin_type)
779 SGEN_ASSERT (0, ((mword)start & (SIZEOF_VOID_P - 1)) == 0, "Why are we scanning for references in unaligned memory ?");
781 #if defined(VALGRIND_MAKE_MEM_DEFINED_IF_ADDRESSABLE) && !defined(_WIN64)
782 VALGRIND_MAKE_MEM_DEFINED_IF_ADDRESSABLE (start, (char*)end - (char*)start);
785 while (start < end) {
787 * *start can point to the middle of an object
788 * note: should we handle pointing at the end of an object?
789 * pinning in C# code disallows pointing at the end of an object
790 * but there is some small chance that an optimizing C compiler
791 * may keep the only reference to an object by pointing
792 * at the end of it. We ignore this small chance for now.
793 * Pointers to the end of an object are indistinguishable
794 * from pointers to the start of the next object in memory
795 * so if we allow that we'd need to pin two objects...
796 * We queue the pointer in an array, the
797 * array will then be sorted and uniqued. This way
798 * we can coalesce several pinning pointers and it should
799 * be faster since we'd do a memory scan with increasing
800 * addresses. Note: we can align the address to the allocation
801 * alignment, so the unique process is more effective.
803 mword addr = (mword)*start;
804 addr &= ~(ALLOC_ALIGN - 1);
805 if (addr >= (mword)start_nursery && addr < (mword)end_nursery) {
806 SGEN_LOG (6, "Pinning address %p from %p", (void*)addr, start);
807 sgen_pin_stage_ptr ((void*)addr);
808 binary_protocol_pin_stage (start, (void*)addr);
809 sgen_pin_stats_register_address ((char*)addr, pin_type);
815 SGEN_LOG (7, "found %d potential pinned heap pointers", count);
819 * The first thing we do in a collection is to identify pinned objects.
820 * This function considers all the areas of memory that need to be
821 * conservatively scanned.
824 pin_from_roots (void *start_nursery, void *end_nursery, ScanCopyContext ctx)
828 SGEN_LOG (2, "Scanning pinned roots (%d bytes, %d/%d entries)", (int)roots_size, roots_hash [ROOT_TYPE_NORMAL].num_entries, roots_hash [ROOT_TYPE_PINNED].num_entries);
829 /* objects pinned from the API are inside these roots */
830 SGEN_HASH_TABLE_FOREACH (&roots_hash [ROOT_TYPE_PINNED], void **, start_root, RootRecord *, root) {
831 SGEN_LOG (6, "Pinned roots %p-%p", start_root, root->end_root);
832 sgen_conservatively_pin_objects_from (start_root, (void**)root->end_root, start_nursery, end_nursery, PIN_TYPE_OTHER);
833 } SGEN_HASH_TABLE_FOREACH_END;
834 /* now deal with the thread stacks
835 * in the future we should be able to conservatively scan only:
836 * *) the cpu registers
837 * *) the unmanaged stack frames
838 * *) the _last_ managed stack frame
839 * *) pointers slots in managed frames
841 sgen_client_scan_thread_data (start_nursery, end_nursery, FALSE, ctx);
845 single_arg_user_copy_or_mark (GCObject **obj, void *gc_data)
847 ScanCopyContext *ctx = (ScanCopyContext *)gc_data;
848 ctx->ops->copy_or_mark_object (obj, ctx->queue);
852 * The memory area from start_root to end_root contains pointers to objects.
853 * Their position is precisely described by @desc (this means that the pointer
854 * can be either NULL or the pointer to the start of an object).
855 * This functions copies them to to_space updates them.
857 * This function is not thread-safe!
860 precisely_scan_objects_from (void** start_root, void** end_root, char* n_start, char *n_end, SgenDescriptor desc, ScanCopyContext ctx)
862 CopyOrMarkObjectFunc copy_func = ctx.ops->copy_or_mark_object;
863 ScanPtrFieldFunc scan_field_func = ctx.ops->scan_ptr_field;
864 SgenGrayQueue *queue = ctx.queue;
866 switch (desc & ROOT_DESC_TYPE_MASK) {
867 case ROOT_DESC_BITMAP:
868 desc >>= ROOT_DESC_TYPE_SHIFT;
870 if ((desc & 1) && *start_root) {
871 copy_func ((GCObject**)start_root, queue);
872 SGEN_LOG (9, "Overwrote root at %p with %p", start_root, *start_root);
878 case ROOT_DESC_COMPLEX: {
879 gsize *bitmap_data = (gsize *)sgen_get_complex_descriptor_bitmap (desc);
880 gsize bwords = (*bitmap_data) - 1;
881 void **start_run = start_root;
883 while (bwords-- > 0) {
884 gsize bmap = *bitmap_data++;
885 void **objptr = start_run;
887 if ((bmap & 1) && *objptr) {
888 copy_func ((GCObject**)objptr, queue);
889 SGEN_LOG (9, "Overwrote root at %p with %p", objptr, *objptr);
894 start_run += GC_BITS_PER_WORD;
898 case ROOT_DESC_VECTOR: {
901 for (p = start_root; p < end_root; p++) {
903 scan_field_func (NULL, (GCObject**)p, queue);
907 case ROOT_DESC_USER: {
908 SgenUserRootMarkFunc marker = sgen_get_user_descriptor_func (desc);
909 marker (start_root, single_arg_user_copy_or_mark, &ctx);
912 case ROOT_DESC_RUN_LEN:
913 g_assert_not_reached ();
915 g_assert_not_reached ();
920 reset_heap_boundaries (void)
922 lowest_heap_address = ~(mword)0;
923 highest_heap_address = 0;
927 sgen_update_heap_boundaries (mword low, mword high)
932 old = lowest_heap_address;
935 } while (SGEN_CAS_PTR ((gpointer*)&lowest_heap_address, (gpointer)low, (gpointer)old) != (gpointer)old);
938 old = highest_heap_address;
941 } while (SGEN_CAS_PTR ((gpointer*)&highest_heap_address, (gpointer)high, (gpointer)old) != (gpointer)old);
945 * Allocate and setup the data structures needed to be able to allocate objects
946 * in the nursery. The nursery is stored in nursery_section.
951 GCMemSection *section;
958 SGEN_LOG (2, "Allocating nursery size: %zu", (size_t)sgen_nursery_size);
959 /* later we will alloc a larger area for the nursery but only activate
960 * what we need. The rest will be used as expansion if we have too many pinned
961 * objects in the existing nursery.
963 /* FIXME: handle OOM */
964 section = (GCMemSection *)sgen_alloc_internal (INTERNAL_MEM_SECTION);
966 alloc_size = sgen_nursery_size;
968 /* If there isn't enough space even for the nursery we should simply abort. */
969 g_assert (sgen_memgov_try_alloc_space (alloc_size, SPACE_NURSERY));
971 data = (char *)major_collector.alloc_heap (alloc_size, alloc_size, DEFAULT_NURSERY_BITS);
972 sgen_update_heap_boundaries ((mword)data, (mword)(data + sgen_nursery_size));
973 SGEN_LOG (4, "Expanding nursery size (%p-%p): %lu, total: %lu", data, data + alloc_size, (unsigned long)sgen_nursery_size, (unsigned long)sgen_gc_get_total_heap_allocation ());
974 section->data = section->next_data = data;
975 section->size = alloc_size;
976 section->end_data = data + sgen_nursery_size;
977 scan_starts = (alloc_size + SCAN_START_SIZE - 1) / SCAN_START_SIZE;
978 section->scan_starts = (char **)sgen_alloc_internal_dynamic (sizeof (char*) * scan_starts, INTERNAL_MEM_SCAN_STARTS, TRUE);
979 section->num_scan_start = scan_starts;
981 nursery_section = section;
983 sgen_nursery_allocator_set_nursery_bounds (data, data + sgen_nursery_size);
987 mono_gc_get_logfile (void)
989 return gc_debug_file;
993 mono_gc_params_set (const char* options)
995 if (gc_params_options)
996 g_free (gc_params_options);
998 gc_params_options = g_strdup (options);
1002 mono_gc_debug_set (const char* options)
1004 if (gc_debug_options)
1005 g_free (gc_debug_options);
1007 gc_debug_options = g_strdup (options);
1011 scan_finalizer_entries (SgenPointerQueue *fin_queue, ScanCopyContext ctx)
1013 CopyOrMarkObjectFunc copy_func = ctx.ops->copy_or_mark_object;
1014 SgenGrayQueue *queue = ctx.queue;
1017 for (i = 0; i < fin_queue->next_slot; ++i) {
1018 GCObject *obj = (GCObject *)fin_queue->data [i];
1021 SGEN_LOG (5, "Scan of fin ready object: %p (%s)\n", obj, sgen_client_vtable_get_name (SGEN_LOAD_VTABLE (obj)));
1022 copy_func ((GCObject**)&fin_queue->data [i], queue);
1027 generation_name (int generation)
1029 switch (generation) {
1030 case GENERATION_NURSERY: return "nursery";
1031 case GENERATION_OLD: return "old";
1032 default: g_assert_not_reached ();
1037 sgen_generation_name (int generation)
1039 return generation_name (generation);
1043 finish_gray_stack (int generation, ScanCopyContext ctx)
1047 int done_with_ephemerons, ephemeron_rounds = 0;
1048 char *start_addr = generation == GENERATION_NURSERY ? sgen_get_nursery_start () : NULL;
1049 char *end_addr = generation == GENERATION_NURSERY ? sgen_get_nursery_end () : (char*)-1;
1050 SgenGrayQueue *queue = ctx.queue;
1052 binary_protocol_finish_gray_stack_start (sgen_timestamp (), generation);
1054 * We copied all the reachable objects. Now it's the time to copy
1055 * the objects that were not referenced by the roots, but by the copied objects.
1056 * we built a stack of objects pointed to by gray_start: they are
1057 * additional roots and we may add more items as we go.
1058 * We loop until gray_start == gray_objects which means no more objects have
1059 * been added. Note this is iterative: no recursion is involved.
1060 * We need to walk the LO list as well in search of marked big objects
1061 * (use a flag since this is needed only on major collections). We need to loop
1062 * here as well, so keep a counter of marked LO (increasing it in copy_object).
1063 * To achieve better cache locality and cache usage, we drain the gray stack
1064 * frequently, after each object is copied, and just finish the work here.
1066 sgen_drain_gray_stack (ctx);
1068 SGEN_LOG (2, "%s generation done", generation_name (generation));
1071 Reset bridge data, we might have lingering data from a previous collection if this is a major
1072 collection trigged by minor overflow.
1074 We must reset the gathered bridges since their original block might be evacuated due to major
1075 fragmentation in the meanwhile and the bridge code should not have to deal with that.
1077 if (sgen_client_bridge_need_processing ())
1078 sgen_client_bridge_reset_data ();
1081 * Mark all strong toggleref objects. This must be done before we walk ephemerons or finalizers
1082 * to ensure they see the full set of live objects.
1084 sgen_client_mark_togglerefs (start_addr, end_addr, ctx);
1087 * Walk the ephemeron tables marking all values with reachable keys. This must be completely done
1088 * before processing finalizable objects and non-tracking weak links to avoid finalizing/clearing
1089 * objects that are in fact reachable.
1091 done_with_ephemerons = 0;
1093 done_with_ephemerons = sgen_client_mark_ephemerons (ctx);
1094 sgen_drain_gray_stack (ctx);
1096 } while (!done_with_ephemerons);
1098 if (sgen_client_bridge_need_processing ()) {
1099 /*Make sure the gray stack is empty before we process bridge objects so we get liveness right*/
1100 sgen_drain_gray_stack (ctx);
1101 sgen_collect_bridge_objects (generation, ctx);
1102 if (generation == GENERATION_OLD)
1103 sgen_collect_bridge_objects (GENERATION_NURSERY, ctx);
1106 Do the first bridge step here, as the collector liveness state will become useless after that.
1108 An important optimization is to only proccess the possibly dead part of the object graph and skip
1109 over all live objects as we transitively know everything they point must be alive too.
1111 The above invariant is completely wrong if we let the gray queue be drained and mark/copy everything.
1113 This has the unfortunate side effect of making overflow collections perform the first step twice, but
1114 given we now have heuristics that perform major GC in anticipation of minor overflows this should not
1117 sgen_client_bridge_processing_stw_step ();
1121 Make sure we drain the gray stack before processing disappearing links and finalizers.
1122 If we don't make sure it is empty we might wrongly see a live object as dead.
1124 sgen_drain_gray_stack (ctx);
1127 We must clear weak links that don't track resurrection before processing object ready for
1128 finalization so they can be cleared before that.
1130 sgen_null_link_in_range (generation, ctx, FALSE);
1131 if (generation == GENERATION_OLD)
1132 sgen_null_link_in_range (GENERATION_NURSERY, ctx, FALSE);
1135 /* walk the finalization queue and move also the objects that need to be
1136 * finalized: use the finalized objects as new roots so the objects they depend
1137 * on are also not reclaimed. As with the roots above, only objects in the nursery
1138 * are marked/copied.
1140 sgen_finalize_in_range (generation, ctx);
1141 if (generation == GENERATION_OLD)
1142 sgen_finalize_in_range (GENERATION_NURSERY, ctx);
1143 /* drain the new stack that might have been created */
1144 SGEN_LOG (6, "Precise scan of gray area post fin");
1145 sgen_drain_gray_stack (ctx);
1148 * This must be done again after processing finalizable objects since CWL slots are cleared only after the key is finalized.
1150 done_with_ephemerons = 0;
1152 done_with_ephemerons = sgen_client_mark_ephemerons (ctx);
1153 sgen_drain_gray_stack (ctx);
1155 } while (!done_with_ephemerons);
1157 sgen_client_clear_unreachable_ephemerons (ctx);
1160 * We clear togglerefs only after all possible chances of revival are done.
1161 * This is semantically more inline with what users expect and it allows for
1162 * user finalizers to correctly interact with TR objects.
1164 sgen_client_clear_togglerefs (start_addr, end_addr, ctx);
1167 SGEN_LOG (2, "Finalize queue handling scan for %s generation: %lld usecs %d ephemeron rounds", generation_name (generation), (long long)TV_ELAPSED (atv, btv), ephemeron_rounds);
1170 * handle disappearing links
1171 * Note we do this after checking the finalization queue because if an object
1172 * survives (at least long enough to be finalized) we don't clear the link.
1173 * This also deals with a possible issue with the monitor reclamation: with the Boehm
1174 * GC a finalized object my lose the monitor because it is cleared before the finalizer is
1177 g_assert (sgen_gray_object_queue_is_empty (queue));
1179 sgen_null_link_in_range (generation, ctx, TRUE);
1180 if (generation == GENERATION_OLD)
1181 sgen_null_link_in_range (GENERATION_NURSERY, ctx, TRUE);
1182 if (sgen_gray_object_queue_is_empty (queue))
1184 sgen_drain_gray_stack (ctx);
1187 g_assert (sgen_gray_object_queue_is_empty (queue));
1189 binary_protocol_finish_gray_stack_end (sgen_timestamp (), generation);
1193 sgen_check_section_scan_starts (GCMemSection *section)
1196 for (i = 0; i < section->num_scan_start; ++i) {
1197 if (section->scan_starts [i]) {
1198 mword size = safe_object_get_size ((GCObject*) section->scan_starts [i]);
1199 SGEN_ASSERT (0, size >= SGEN_CLIENT_MINIMUM_OBJECT_SIZE && size <= MAX_SMALL_OBJ_SIZE, "Weird object size at scan starts.");
1205 check_scan_starts (void)
1207 if (!do_scan_starts_check)
1209 sgen_check_section_scan_starts (nursery_section);
1210 major_collector.check_scan_starts ();
1214 scan_from_registered_roots (char *addr_start, char *addr_end, int root_type, ScanCopyContext ctx)
1218 SGEN_HASH_TABLE_FOREACH (&roots_hash [root_type], void **, start_root, RootRecord *, root) {
1219 SGEN_LOG (6, "Precise root scan %p-%p (desc: %p)", start_root, root->end_root, (void*)root->root_desc);
1220 precisely_scan_objects_from (start_root, (void**)root->end_root, addr_start, addr_end, root->root_desc, ctx);
1221 } SGEN_HASH_TABLE_FOREACH_END;
1227 static gboolean inited = FALSE;
1232 mono_counters_register ("Collection max time", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME | MONO_COUNTER_MONOTONIC, &time_max);
1234 mono_counters_register ("Minor fragment clear", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_pre_collection_fragment_clear);
1235 mono_counters_register ("Minor pinning", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_pinning);
1236 mono_counters_register ("Minor scan remembered set", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_scan_remsets);
1237 mono_counters_register ("Minor scan major blocks", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_scan_major_blocks);
1238 mono_counters_register ("Minor scan los", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_scan_los);
1239 mono_counters_register ("Minor scan pinned", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_scan_pinned);
1240 mono_counters_register ("Minor scan roots", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_scan_roots);
1241 mono_counters_register ("Minor fragment creation", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_fragment_creation);
1243 mono_counters_register ("Major fragment clear", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_pre_collection_fragment_clear);
1244 mono_counters_register ("Major pinning", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_pinning);
1245 mono_counters_register ("Major scan pinned", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_scan_pinned);
1246 mono_counters_register ("Major scan roots", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_scan_roots);
1247 mono_counters_register ("Major scan mod union", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_scan_mod_union);
1248 mono_counters_register ("Major finish gray stack", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_finish_gray_stack);
1249 mono_counters_register ("Major free big objects", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_free_bigobjs);
1250 mono_counters_register ("Major LOS sweep", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_los_sweep);
1251 mono_counters_register ("Major sweep", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_sweep);
1252 mono_counters_register ("Major fragment creation", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_fragment_creation);
1254 mono_counters_register ("Number of pinned objects", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_pinned_objects);
1256 #ifdef HEAVY_STATISTICS
1257 mono_counters_register ("WBarrier remember pointer", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_wbarrier_add_to_global_remset);
1258 mono_counters_register ("WBarrier arrayref copy", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_wbarrier_arrayref_copy);
1259 mono_counters_register ("WBarrier generic store called", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_wbarrier_generic_store);
1260 mono_counters_register ("WBarrier generic atomic store called", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_wbarrier_generic_store_atomic);
1261 mono_counters_register ("WBarrier set root", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_wbarrier_set_root);
1263 mono_counters_register ("# objects allocated degraded", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_objects_alloced_degraded);
1264 mono_counters_register ("bytes allocated degraded", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_bytes_alloced_degraded);
1266 mono_counters_register ("# copy_object() called (nursery)", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_copy_object_called_nursery);
1267 mono_counters_register ("# objects copied (nursery)", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_objects_copied_nursery);
1268 mono_counters_register ("# copy_object() called (major)", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_copy_object_called_major);
1269 mono_counters_register ("# objects copied (major)", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_objects_copied_major);
1271 mono_counters_register ("# scan_object() called (nursery)", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_scan_object_called_nursery);
1272 mono_counters_register ("# scan_object() called (major)", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_scan_object_called_major);
1274 mono_counters_register ("Slots allocated in vain", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_slots_allocated_in_vain);
1276 mono_counters_register ("# nursery copy_object() failed from space", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_nursery_copy_object_failed_from_space);
1277 mono_counters_register ("# nursery copy_object() failed forwarded", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_nursery_copy_object_failed_forwarded);
1278 mono_counters_register ("# nursery copy_object() failed pinned", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_nursery_copy_object_failed_pinned);
1279 mono_counters_register ("# nursery copy_object() failed to space", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_nursery_copy_object_failed_to_space);
1281 sgen_nursery_allocator_init_heavy_stats ();
1289 reset_pinned_from_failed_allocation (void)
1291 bytes_pinned_from_failed_allocation = 0;
1295 sgen_set_pinned_from_failed_allocation (mword objsize)
1297 bytes_pinned_from_failed_allocation += objsize;
1301 sgen_collection_is_concurrent (void)
1303 switch (current_collection_generation) {
1304 case GENERATION_NURSERY:
1306 case GENERATION_OLD:
1307 return concurrent_collection_in_progress;
1309 g_error ("Invalid current generation %d", current_collection_generation);
1315 sgen_concurrent_collection_in_progress (void)
1317 return concurrent_collection_in_progress;
1321 SgenThreadPoolJob job;
1322 SgenObjectOperations *ops;
1323 SgenGrayQueue *gc_thread_gray_queue;
1331 static ScanCopyContext
1332 scan_copy_context_for_scan_job (void *worker_data_untyped, ScanJob *job)
1334 WorkerData *worker_data = (WorkerData *)worker_data_untyped;
1336 return CONTEXT_FROM_OBJECT_OPERATIONS (job->ops, sgen_workers_get_job_gray_queue (worker_data, job->gc_thread_gray_queue));
1344 } ScanFromRegisteredRootsJob;
1347 job_scan_from_registered_roots (void *worker_data_untyped, SgenThreadPoolJob *job)
1349 ScanFromRegisteredRootsJob *job_data = (ScanFromRegisteredRootsJob*)job;
1350 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, &job_data->scan_job);
1352 scan_from_registered_roots (job_data->heap_start, job_data->heap_end, job_data->root_type, ctx);
1359 } ScanThreadDataJob;
1362 job_scan_thread_data (void *worker_data_untyped, SgenThreadPoolJob *job)
1364 ScanThreadDataJob *job_data = (ScanThreadDataJob*)job;
1365 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, &job_data->scan_job);
1367 sgen_client_scan_thread_data (job_data->heap_start, job_data->heap_end, TRUE, ctx);
1372 SgenPointerQueue *queue;
1373 } ScanFinalizerEntriesJob;
1376 job_scan_finalizer_entries (void *worker_data_untyped, SgenThreadPoolJob *job)
1378 ScanFinalizerEntriesJob *job_data = (ScanFinalizerEntriesJob*)job;
1379 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, &job_data->scan_job);
1381 scan_finalizer_entries (job_data->queue, ctx);
1385 job_scan_wbroots (void *worker_data_untyped, SgenThreadPoolJob *job)
1387 ScanJob *job_data = (ScanJob*)job;
1388 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, job_data);
1390 sgen_wbroots_scan_card_table (ctx);
1394 job_scan_major_card_table (void *worker_data_untyped, SgenThreadPoolJob *job)
1396 SGEN_TV_DECLARE (atv);
1397 SGEN_TV_DECLARE (btv);
1398 ParallelScanJob *job_data = (ParallelScanJob*)job;
1399 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, (ScanJob*)job_data);
1401 SGEN_TV_GETTIME (atv);
1402 major_collector.scan_card_table (CARDTABLE_SCAN_GLOBAL, ctx, job_data->job_index, sgen_workers_get_job_split_count ());
1403 SGEN_TV_GETTIME (btv);
1404 time_minor_scan_major_blocks += SGEN_TV_ELAPSED (atv, btv);
1408 job_scan_los_card_table (void *worker_data_untyped, SgenThreadPoolJob *job)
1410 SGEN_TV_DECLARE (atv);
1411 SGEN_TV_DECLARE (btv);
1412 ParallelScanJob *job_data = (ParallelScanJob*)job;
1413 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, (ScanJob*)job_data);
1415 SGEN_TV_GETTIME (atv);
1416 sgen_los_scan_card_table (CARDTABLE_SCAN_GLOBAL, ctx, job_data->job_index, sgen_workers_get_job_split_count ());
1417 SGEN_TV_GETTIME (btv);
1418 time_minor_scan_los += SGEN_TV_ELAPSED (atv, btv);
1422 job_scan_major_mod_union_card_table (void *worker_data_untyped, SgenThreadPoolJob *job)
1424 ParallelScanJob *job_data = (ParallelScanJob*)job;
1425 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, (ScanJob*)job_data);
1427 g_assert (concurrent_collection_in_progress);
1428 major_collector.scan_card_table (CARDTABLE_SCAN_MOD_UNION, ctx, job_data->job_index, sgen_workers_get_job_split_count ());
1432 job_scan_los_mod_union_card_table (void *worker_data_untyped, SgenThreadPoolJob *job)
1434 ParallelScanJob *job_data = (ParallelScanJob*)job;
1435 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, (ScanJob*)job_data);
1437 g_assert (concurrent_collection_in_progress);
1438 sgen_los_scan_card_table (CARDTABLE_SCAN_MOD_UNION, ctx, job_data->job_index, sgen_workers_get_job_split_count ());
1442 job_major_mod_union_preclean (void *worker_data_untyped, SgenThreadPoolJob *job)
1444 ParallelScanJob *job_data = (ParallelScanJob*)job;
1445 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, (ScanJob*)job_data);
1447 g_assert (concurrent_collection_in_progress);
1449 major_collector.scan_card_table (CARDTABLE_SCAN_MOD_UNION_PRECLEAN, ctx, job_data->job_index, sgen_workers_get_job_split_count ());
1453 job_los_mod_union_preclean (void *worker_data_untyped, SgenThreadPoolJob *job)
1455 ParallelScanJob *job_data = (ParallelScanJob*)job;
1456 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, (ScanJob*)job_data);
1458 g_assert (concurrent_collection_in_progress);
1460 sgen_los_scan_card_table (CARDTABLE_SCAN_MOD_UNION_PRECLEAN, ctx, job_data->job_index, sgen_workers_get_job_split_count ());
1464 job_scan_last_pinned (void *worker_data_untyped, SgenThreadPoolJob *job)
1466 ScanJob *job_data = (ScanJob*)job;
1467 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, job_data);
1469 g_assert (concurrent_collection_in_progress);
1471 sgen_scan_pin_queue_objects (ctx);
1475 workers_finish_callback (void)
1477 ParallelScanJob *psj;
1479 int split_count = sgen_workers_get_job_split_count ();
1481 /* Mod union preclean jobs */
1482 for (i = 0; i < split_count; i++) {
1483 psj = (ParallelScanJob*)sgen_thread_pool_job_alloc ("preclean major mod union cardtable", job_major_mod_union_preclean, sizeof (ParallelScanJob));
1484 psj->scan_job.ops = sgen_workers_get_idle_func_object_ops ();
1485 psj->scan_job.gc_thread_gray_queue = NULL;
1487 sgen_workers_enqueue_job (&psj->scan_job.job, TRUE);
1490 for (i = 0; i < split_count; i++) {
1491 psj = (ParallelScanJob*)sgen_thread_pool_job_alloc ("preclean los mod union cardtable", job_los_mod_union_preclean, sizeof (ParallelScanJob));
1492 psj->scan_job.ops = sgen_workers_get_idle_func_object_ops ();
1493 psj->scan_job.gc_thread_gray_queue = NULL;
1495 sgen_workers_enqueue_job (&psj->scan_job.job, TRUE);
1498 sj = (ScanJob*)sgen_thread_pool_job_alloc ("scan last pinned", job_scan_last_pinned, sizeof (ScanJob));
1499 sj->ops = sgen_workers_get_idle_func_object_ops ();
1500 sj->gc_thread_gray_queue = NULL;
1501 sgen_workers_enqueue_job (&sj->job, TRUE);
1505 init_gray_queue (SgenGrayQueue *gc_thread_gray_queue, gboolean use_workers)
1508 sgen_workers_init_distribute_gray_queue ();
1509 sgen_gray_object_queue_init (gc_thread_gray_queue, NULL, TRUE);
1513 enqueue_scan_remembered_set_jobs (SgenGrayQueue *gc_thread_gray_queue, SgenObjectOperations *ops, gboolean enqueue)
1515 int i, split_count = sgen_workers_get_job_split_count ();
1518 sj = (ScanJob*)sgen_thread_pool_job_alloc ("scan wbroots", job_scan_wbroots, sizeof (ScanJob));
1520 sj->gc_thread_gray_queue = gc_thread_gray_queue;
1521 sgen_workers_enqueue_job (&sj->job, enqueue);
1523 for (i = 0; i < split_count; i++) {
1524 ParallelScanJob *psj;
1526 psj = (ParallelScanJob*)sgen_thread_pool_job_alloc ("scan major remsets", job_scan_major_card_table, sizeof (ParallelScanJob));
1527 psj->scan_job.ops = ops;
1528 psj->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1530 sgen_workers_enqueue_job (&psj->scan_job.job, enqueue);
1532 psj = (ParallelScanJob*)sgen_thread_pool_job_alloc ("scan LOS remsets", job_scan_los_card_table, sizeof (ParallelScanJob));
1533 psj->scan_job.ops = ops;
1534 psj->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1536 sgen_workers_enqueue_job (&psj->scan_job.job, enqueue);
1541 enqueue_scan_from_roots_jobs (SgenGrayQueue *gc_thread_gray_queue, char *heap_start, char *heap_end, SgenObjectOperations *ops, gboolean enqueue)
1543 ScanFromRegisteredRootsJob *scrrj;
1544 ScanThreadDataJob *stdj;
1545 ScanFinalizerEntriesJob *sfej;
1547 /* registered roots, this includes static fields */
1549 scrrj = (ScanFromRegisteredRootsJob*)sgen_thread_pool_job_alloc ("scan from registered roots normal", job_scan_from_registered_roots, sizeof (ScanFromRegisteredRootsJob));
1550 scrrj->scan_job.ops = ops;
1551 scrrj->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1552 scrrj->heap_start = heap_start;
1553 scrrj->heap_end = heap_end;
1554 scrrj->root_type = ROOT_TYPE_NORMAL;
1555 sgen_workers_enqueue_job (&scrrj->scan_job.job, enqueue);
1557 if (current_collection_generation == GENERATION_OLD) {
1558 /* During minors we scan the cardtable for these roots instead */
1559 scrrj = (ScanFromRegisteredRootsJob*)sgen_thread_pool_job_alloc ("scan from registered roots wbarrier", job_scan_from_registered_roots, sizeof (ScanFromRegisteredRootsJob));
1560 scrrj->scan_job.ops = ops;
1561 scrrj->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1562 scrrj->heap_start = heap_start;
1563 scrrj->heap_end = heap_end;
1564 scrrj->root_type = ROOT_TYPE_WBARRIER;
1565 sgen_workers_enqueue_job (&scrrj->scan_job.job, enqueue);
1570 stdj = (ScanThreadDataJob*)sgen_thread_pool_job_alloc ("scan thread data", job_scan_thread_data, sizeof (ScanThreadDataJob));
1571 stdj->scan_job.ops = ops;
1572 stdj->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1573 stdj->heap_start = heap_start;
1574 stdj->heap_end = heap_end;
1575 sgen_workers_enqueue_job (&stdj->scan_job.job, enqueue);
1577 /* Scan the list of objects ready for finalization. */
1579 sfej = (ScanFinalizerEntriesJob*)sgen_thread_pool_job_alloc ("scan finalizer entries", job_scan_finalizer_entries, sizeof (ScanFinalizerEntriesJob));
1580 sfej->scan_job.ops = ops;
1581 sfej->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1582 sfej->queue = &fin_ready_queue;
1583 sgen_workers_enqueue_job (&sfej->scan_job.job, enqueue);
1585 sfej = (ScanFinalizerEntriesJob*)sgen_thread_pool_job_alloc ("scan critical finalizer entries", job_scan_finalizer_entries, sizeof (ScanFinalizerEntriesJob));
1586 sfej->scan_job.ops = ops;
1587 sfej->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1588 sfej->queue = &critical_fin_queue;
1589 sgen_workers_enqueue_job (&sfej->scan_job.job, enqueue);
1593 * Perform a nursery collection.
1595 * Return whether any objects were late-pinned due to being out of memory.
1598 collect_nursery (const char *reason, gboolean is_overflow, SgenGrayQueue *unpin_queue)
1600 gboolean needs_major, is_parallel = FALSE;
1601 size_t max_garbage_amount;
1603 mword fragment_total;
1604 SgenGrayQueue gc_thread_gray_queue;
1605 SgenObjectOperations *object_ops_nopar, *object_ops_par = NULL;
1606 ScanCopyContext ctx;
1609 SGEN_TV_DECLARE (last_minor_collection_start_tv);
1610 SGEN_TV_DECLARE (last_minor_collection_end_tv);
1612 if (disable_minor_collections)
1615 TV_GETTIME (last_minor_collection_start_tv);
1616 atv = last_minor_collection_start_tv;
1618 binary_protocol_collection_begin (gc_stats.minor_gc_count, GENERATION_NURSERY);
1620 if (sgen_concurrent_collection_in_progress ()) {
1621 /* FIXME Support parallel nursery collections with concurrent major */
1622 object_ops_nopar = &sgen_minor_collector.serial_ops_with_concurrent_major;
1624 object_ops_nopar = &sgen_minor_collector.serial_ops;
1625 if (sgen_minor_collector.is_parallel) {
1626 object_ops_par = &sgen_minor_collector.parallel_ops;
1631 if (do_verify_nursery || do_dump_nursery_content)
1632 sgen_debug_verify_nursery (do_dump_nursery_content);
1634 current_collection_generation = GENERATION_NURSERY;
1636 SGEN_ASSERT (0, !sgen_collection_is_concurrent (), "Why is the nursery collection concurrent?");
1638 reset_pinned_from_failed_allocation ();
1640 check_scan_starts ();
1642 sgen_nursery_alloc_prepare_for_minor ();
1646 nursery_next = sgen_nursery_alloc_get_upper_alloc_bound ();
1647 /* FIXME: optimize later to use the higher address where an object can be present */
1648 nursery_next = MAX (nursery_next, sgen_get_nursery_end ());
1650 SGEN_LOG (1, "Start nursery collection %d %p-%p, size: %d", gc_stats.minor_gc_count, sgen_get_nursery_start (), nursery_next, (int)(nursery_next - sgen_get_nursery_start ()));
1651 max_garbage_amount = nursery_next - sgen_get_nursery_start ();
1652 g_assert (nursery_section->size >= max_garbage_amount);
1654 /* world must be stopped already */
1656 time_minor_pre_collection_fragment_clear += TV_ELAPSED (atv, btv);
1658 sgen_client_pre_collection_checks ();
1660 nursery_section->next_data = nursery_next;
1662 major_collector.start_nursery_collection ();
1664 sgen_memgov_minor_collection_start ();
1666 init_gray_queue (&gc_thread_gray_queue, is_parallel);
1667 ctx = CONTEXT_FROM_OBJECT_OPERATIONS (object_ops_nopar, &gc_thread_gray_queue);
1669 gc_stats.minor_gc_count ++;
1671 sgen_process_fin_stage_entries ();
1673 /* pin from pinned handles */
1674 sgen_init_pinning ();
1675 sgen_client_binary_protocol_mark_start (GENERATION_NURSERY);
1676 pin_from_roots (sgen_get_nursery_start (), nursery_next, ctx);
1677 /* pin cemented objects */
1678 sgen_pin_cemented_objects ();
1679 /* identify pinned objects */
1680 sgen_optimize_pin_queue ();
1681 sgen_pinning_setup_section (nursery_section);
1683 pin_objects_in_nursery (FALSE, ctx);
1684 sgen_pinning_trim_queue_to_section (nursery_section);
1686 if (remset_consistency_checks)
1687 sgen_check_remset_consistency ();
1689 if (whole_heap_check_before_collection) {
1690 sgen_clear_nursery_fragments ();
1691 sgen_check_whole_heap (FALSE);
1695 time_minor_pinning += TV_ELAPSED (btv, atv);
1696 SGEN_LOG (2, "Finding pinned pointers: %zd in %lld usecs", sgen_get_pinned_count (), (long long)TV_ELAPSED (btv, atv));
1697 SGEN_LOG (4, "Start scan with %zd pinned objects", sgen_get_pinned_count ());
1699 remset.start_scan_remsets ();
1701 enqueue_scan_remembered_set_jobs (&gc_thread_gray_queue, is_parallel ? object_ops_par : object_ops_nopar, is_parallel);
1703 /* we don't have complete write barrier yet, so we scan all the old generation sections */
1705 time_minor_scan_remsets += TV_ELAPSED (atv, btv);
1706 SGEN_LOG (2, "Old generation scan: %lld usecs", (long long)TV_ELAPSED (atv, btv));
1708 sgen_pin_stats_report ();
1710 /* FIXME: Why do we do this at this specific, seemingly random, point? */
1711 sgen_client_collecting_minor (&fin_ready_queue, &critical_fin_queue);
1714 time_minor_scan_pinned += TV_ELAPSED (btv, atv);
1716 enqueue_scan_from_roots_jobs (&gc_thread_gray_queue, sgen_get_nursery_start (), nursery_next, is_parallel ? object_ops_par : object_ops_nopar, is_parallel);
1719 gray_queue_redirect (&gc_thread_gray_queue);
1720 sgen_workers_start_all_workers (object_ops_nopar, object_ops_par, NULL);
1721 sgen_workers_join ();
1725 time_minor_scan_roots += TV_ELAPSED (atv, btv);
1727 finish_gray_stack (GENERATION_NURSERY, ctx);
1730 time_minor_finish_gray_stack += TV_ELAPSED (btv, atv);
1731 sgen_client_binary_protocol_mark_end (GENERATION_NURSERY);
1733 if (objects_pinned) {
1734 sgen_optimize_pin_queue ();
1735 sgen_pinning_setup_section (nursery_section);
1739 * This is the latest point at which we can do this check, because
1740 * sgen_build_nursery_fragments() unpins nursery objects again.
1742 if (remset_consistency_checks)
1743 sgen_check_remset_consistency ();
1745 /* walk the pin_queue, build up the fragment list of free memory, unmark
1746 * pinned objects as we go, memzero() the empty fragments so they are ready for the
1749 sgen_client_binary_protocol_reclaim_start (GENERATION_NURSERY);
1750 fragment_total = sgen_build_nursery_fragments (nursery_section, unpin_queue);
1751 if (!fragment_total)
1754 /* Clear TLABs for all threads */
1755 sgen_clear_tlabs ();
1757 sgen_client_binary_protocol_reclaim_end (GENERATION_NURSERY);
1759 time_minor_fragment_creation += TV_ELAPSED (atv, btv);
1760 SGEN_LOG (2, "Fragment creation: %lld usecs, %lu bytes available", (long long)TV_ELAPSED (atv, btv), (unsigned long)fragment_total);
1762 if (remset_consistency_checks)
1763 sgen_check_major_refs ();
1765 major_collector.finish_nursery_collection ();
1767 TV_GETTIME (last_minor_collection_end_tv);
1768 gc_stats.minor_gc_time += TV_ELAPSED (last_minor_collection_start_tv, last_minor_collection_end_tv);
1770 sgen_debug_dump_heap ("minor", gc_stats.minor_gc_count - 1, NULL);
1772 /* prepare the pin queue for the next collection */
1773 sgen_finish_pinning ();
1774 if (sgen_have_pending_finalizers ()) {
1775 SGEN_LOG (4, "Finalizer-thread wakeup");
1776 sgen_client_finalize_notify ();
1778 sgen_pin_stats_reset ();
1779 /* clear cemented hash */
1780 sgen_cement_clear_below_threshold ();
1782 sgen_gray_object_queue_dispose (&gc_thread_gray_queue);
1784 check_scan_starts ();
1786 binary_protocol_flush_buffers (FALSE);
1788 sgen_memgov_minor_collection_end (reason, is_overflow);
1790 /*objects are late pinned because of lack of memory, so a major is a good call*/
1791 needs_major = objects_pinned > 0;
1792 current_collection_generation = -1;
1795 binary_protocol_collection_end (gc_stats.minor_gc_count - 1, GENERATION_NURSERY, 0, 0);
1797 if (check_nursery_objects_pinned && !sgen_minor_collector.is_split)
1798 sgen_check_nursery_objects_pinned (unpin_queue != NULL);
1804 COPY_OR_MARK_FROM_ROOTS_SERIAL,
1805 COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT,
1806 COPY_OR_MARK_FROM_ROOTS_FINISH_CONCURRENT
1807 } CopyOrMarkFromRootsMode;
1810 major_copy_or_mark_from_roots (SgenGrayQueue *gc_thread_gray_queue, size_t *old_next_pin_slot, CopyOrMarkFromRootsMode mode, SgenObjectOperations *object_ops_nopar, SgenObjectOperations *object_ops_par)
1815 /* FIXME: only use these values for the precise scan
1816 * note that to_space pointers should be excluded anyway...
1818 char *heap_start = NULL;
1819 char *heap_end = (char*)-1;
1820 ScanCopyContext ctx = CONTEXT_FROM_OBJECT_OPERATIONS (object_ops_nopar, gc_thread_gray_queue);
1821 gboolean concurrent = mode != COPY_OR_MARK_FROM_ROOTS_SERIAL;
1823 SGEN_ASSERT (0, !!concurrent == !!concurrent_collection_in_progress, "We've been called with the wrong mode.");
1825 if (mode == COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT) {
1826 /*This cleans up unused fragments */
1827 sgen_nursery_allocator_prepare_for_pinning ();
1829 if (do_concurrent_checks)
1830 sgen_debug_check_nursery_is_clean ();
1832 /* The concurrent collector doesn't touch the nursery. */
1833 sgen_nursery_alloc_prepare_for_major ();
1838 /* Pinning depends on this */
1839 sgen_clear_nursery_fragments ();
1841 if (whole_heap_check_before_collection)
1842 sgen_check_whole_heap (TRUE);
1845 time_major_pre_collection_fragment_clear += TV_ELAPSED (atv, btv);
1847 if (!sgen_collection_is_concurrent ())
1848 nursery_section->next_data = sgen_get_nursery_end ();
1849 /* we should also coalesce scanning from sections close to each other
1850 * and deal with pointers outside of the sections later.
1855 sgen_client_pre_collection_checks ();
1857 if (mode != COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT) {
1858 /* Remsets are not useful for a major collection */
1859 remset.clear_cards ();
1862 sgen_process_fin_stage_entries ();
1865 sgen_init_pinning ();
1866 SGEN_LOG (6, "Collecting pinned addresses");
1867 pin_from_roots ((void*)lowest_heap_address, (void*)highest_heap_address, ctx);
1868 if (mode == COPY_OR_MARK_FROM_ROOTS_FINISH_CONCURRENT) {
1869 /* Pin cemented objects that were forced */
1870 sgen_pin_cemented_objects ();
1872 sgen_optimize_pin_queue ();
1873 if (mode == COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT) {
1875 * Cemented objects that are in the pinned list will be marked. When
1876 * marking concurrently we won't mark mod-union cards for these objects.
1877 * Instead they will remain cemented until the next major collection,
1878 * when we will recheck if they are still pinned in the roots.
1880 sgen_cement_force_pinned ();
1883 sgen_client_collecting_major_1 ();
1886 * pin_queue now contains all candidate pointers, sorted and
1887 * uniqued. We must do two passes now to figure out which
1888 * objects are pinned.
1890 * The first is to find within the pin_queue the area for each
1891 * section. This requires that the pin_queue be sorted. We
1892 * also process the LOS objects and pinned chunks here.
1894 * The second, destructive, pass is to reduce the section
1895 * areas to pointers to the actually pinned objects.
1897 SGEN_LOG (6, "Pinning from sections");
1898 /* first pass for the sections */
1899 sgen_find_section_pin_queue_start_end (nursery_section);
1900 /* identify possible pointers to the insize of large objects */
1901 SGEN_LOG (6, "Pinning from large objects");
1902 for (bigobj = los_object_list; bigobj; bigobj = bigobj->next) {
1904 if (sgen_find_optimized_pin_queue_area ((char*)bigobj->data, (char*)bigobj->data + sgen_los_object_size (bigobj), &dummy, &dummy)) {
1905 binary_protocol_pin (bigobj->data, (gpointer)LOAD_VTABLE (bigobj->data), safe_object_get_size (bigobj->data));
1907 if (sgen_los_object_is_pinned (bigobj->data)) {
1908 SGEN_ASSERT (0, mode == COPY_OR_MARK_FROM_ROOTS_FINISH_CONCURRENT, "LOS objects can only be pinned here after concurrent marking.");
1911 sgen_los_pin_object (bigobj->data);
1912 if (SGEN_OBJECT_HAS_REFERENCES (bigobj->data))
1913 GRAY_OBJECT_ENQUEUE_SERIAL (gc_thread_gray_queue, bigobj->data, sgen_obj_get_descriptor ((GCObject*)bigobj->data));
1914 sgen_pin_stats_register_object (bigobj->data, GENERATION_OLD);
1915 SGEN_LOG (6, "Marked large object %p (%s) size: %lu from roots", bigobj->data,
1916 sgen_client_vtable_get_name (SGEN_LOAD_VTABLE (bigobj->data)),
1917 (unsigned long)sgen_los_object_size (bigobj));
1919 sgen_client_pinned_los_object (bigobj->data);
1923 pin_objects_in_nursery (mode == COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT, ctx);
1924 if (check_nursery_objects_pinned && !sgen_minor_collector.is_split)
1925 sgen_check_nursery_objects_pinned (mode != COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT);
1927 major_collector.pin_objects (gc_thread_gray_queue);
1928 if (old_next_pin_slot)
1929 *old_next_pin_slot = sgen_get_pinned_count ();
1932 time_major_pinning += TV_ELAPSED (atv, btv);
1933 SGEN_LOG (2, "Finding pinned pointers: %zd in %lld usecs", sgen_get_pinned_count (), (long long)TV_ELAPSED (atv, btv));
1934 SGEN_LOG (4, "Start scan with %zd pinned objects", sgen_get_pinned_count ());
1936 major_collector.init_to_space ();
1938 SGEN_ASSERT (0, sgen_workers_all_done (), "Why are the workers not done when we start or finish a major collection?");
1939 if (mode == COPY_OR_MARK_FROM_ROOTS_FINISH_CONCURRENT) {
1940 if (object_ops_par != NULL)
1941 sgen_workers_set_num_active_workers (0);
1942 if (sgen_workers_have_idle_work ()) {
1944 * We force the finish of the worker with the new object ops context
1945 * which can also do copying. We need to have finished pinning.
1947 sgen_workers_start_all_workers (object_ops_nopar, object_ops_par, NULL);
1949 sgen_workers_join ();
1953 #ifdef SGEN_DEBUG_INTERNAL_ALLOC
1954 main_gc_thread = mono_native_thread_self ();
1957 sgen_client_collecting_major_2 ();
1960 time_major_scan_pinned += TV_ELAPSED (btv, atv);
1962 sgen_client_collecting_major_3 (&fin_ready_queue, &critical_fin_queue);
1964 enqueue_scan_from_roots_jobs (gc_thread_gray_queue, heap_start, heap_end, object_ops_nopar, FALSE);
1967 time_major_scan_roots += TV_ELAPSED (atv, btv);
1970 * We start the concurrent worker after pinning and after we scanned the roots
1971 * in order to make sure that the worker does not finish before handling all
1974 if (mode == COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT) {
1975 sgen_workers_set_num_active_workers (1);
1976 gray_queue_redirect (gc_thread_gray_queue);
1977 if (precleaning_enabled) {
1978 sgen_workers_start_all_workers (object_ops_nopar, object_ops_par, workers_finish_callback);
1980 sgen_workers_start_all_workers (object_ops_nopar, object_ops_par, NULL);
1984 if (mode == COPY_OR_MARK_FROM_ROOTS_FINISH_CONCURRENT) {
1985 int i, split_count = sgen_workers_get_job_split_count ();
1986 gboolean parallel = object_ops_par != NULL;
1988 /* If we're not parallel we finish the collection on the gc thread */
1990 gray_queue_redirect (gc_thread_gray_queue);
1992 /* Mod union card table */
1993 for (i = 0; i < split_count; i++) {
1994 ParallelScanJob *psj;
1996 psj = (ParallelScanJob*)sgen_thread_pool_job_alloc ("scan mod union cardtable", job_scan_major_mod_union_card_table, sizeof (ParallelScanJob));
1997 psj->scan_job.ops = object_ops_par ? object_ops_par : object_ops_nopar;
1998 psj->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
2000 sgen_workers_enqueue_job (&psj->scan_job.job, parallel);
2002 psj = (ParallelScanJob*)sgen_thread_pool_job_alloc ("scan LOS mod union cardtable", job_scan_los_mod_union_card_table, sizeof (ParallelScanJob));
2003 psj->scan_job.ops = object_ops_par ? object_ops_par : object_ops_nopar;
2004 psj->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
2006 sgen_workers_enqueue_job (&psj->scan_job.job, parallel);
2011 * If we enqueue a job while workers are running we need to sgen_workers_ensure_awake
2012 * in order to make sure that we are running the idle func and draining all worker
2013 * gray queues. The operation of starting workers implies this, so we start them after
2014 * in order to avoid doing this operation twice. The workers will drain the main gray
2015 * stack that contained roots and pinned objects and also scan the mod union card
2018 sgen_workers_start_all_workers (object_ops_nopar, object_ops_par, NULL);
2019 sgen_workers_join ();
2023 sgen_pin_stats_report ();
2025 if (mode == COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT) {
2026 sgen_finish_pinning ();
2028 sgen_pin_stats_reset ();
2030 if (do_concurrent_checks)
2031 sgen_debug_check_nursery_is_clean ();
2036 major_start_collection (SgenGrayQueue *gc_thread_gray_queue, const char *reason, gboolean concurrent, size_t *old_next_pin_slot)
2038 SgenObjectOperations *object_ops_nopar, *object_ops_par = NULL;
2040 binary_protocol_collection_begin (gc_stats.major_gc_count, GENERATION_OLD);
2042 current_collection_generation = GENERATION_OLD;
2044 sgen_workers_assert_gray_queue_is_empty ();
2047 sgen_cement_reset ();
2050 g_assert (major_collector.is_concurrent);
2051 concurrent_collection_in_progress = TRUE;
2053 object_ops_nopar = &major_collector.major_ops_concurrent_start;
2054 if (major_collector.is_parallel)
2055 object_ops_par = &major_collector.major_ops_conc_par_start;
2058 object_ops_nopar = &major_collector.major_ops_serial;
2061 reset_pinned_from_failed_allocation ();
2063 sgen_memgov_major_collection_start (concurrent, reason);
2065 //count_ref_nonref_objs ();
2066 //consistency_check ();
2068 check_scan_starts ();
2071 SGEN_LOG (1, "Start major collection %d", gc_stats.major_gc_count);
2072 gc_stats.major_gc_count ++;
2074 if (major_collector.start_major_collection)
2075 major_collector.start_major_collection ();
2077 major_copy_or_mark_from_roots (gc_thread_gray_queue, old_next_pin_slot, concurrent ? COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT : COPY_OR_MARK_FROM_ROOTS_SERIAL, object_ops_nopar, object_ops_par);
2081 major_finish_collection (SgenGrayQueue *gc_thread_gray_queue, const char *reason, gboolean is_overflow, size_t old_next_pin_slot, gboolean forced)
2083 ScannedObjectCounts counts;
2084 SgenObjectOperations *object_ops_nopar;
2085 mword fragment_total;
2091 if (concurrent_collection_in_progress) {
2092 SgenObjectOperations *object_ops_par = NULL;
2094 object_ops_nopar = &major_collector.major_ops_concurrent_finish;
2095 if (major_collector.is_parallel)
2096 object_ops_par = &major_collector.major_ops_conc_par_finish;
2098 major_copy_or_mark_from_roots (gc_thread_gray_queue, NULL, COPY_OR_MARK_FROM_ROOTS_FINISH_CONCURRENT, object_ops_nopar, object_ops_par);
2100 #ifdef SGEN_DEBUG_INTERNAL_ALLOC
2101 main_gc_thread = NULL;
2104 object_ops_nopar = &major_collector.major_ops_serial;
2107 sgen_workers_assert_gray_queue_is_empty ();
2109 finish_gray_stack (GENERATION_OLD, CONTEXT_FROM_OBJECT_OPERATIONS (object_ops_nopar, gc_thread_gray_queue));
2111 time_major_finish_gray_stack += TV_ELAPSED (btv, atv);
2113 SGEN_ASSERT (0, sgen_workers_all_done (), "Can't have workers working after joining");
2115 if (objects_pinned) {
2116 g_assert (!concurrent_collection_in_progress);
2119 * This is slow, but we just OOM'd.
2121 * See comment at `sgen_pin_queue_clear_discarded_entries` for how the pin
2122 * queue is laid out at this point.
2124 sgen_pin_queue_clear_discarded_entries (nursery_section, old_next_pin_slot);
2126 * We need to reestablish all pinned nursery objects in the pin queue
2127 * because they're needed for fragment creation. Unpinning happens by
2128 * walking the whole queue, so it's not necessary to reestablish where major
2129 * heap block pins are - all we care is that they're still in there
2132 sgen_optimize_pin_queue ();
2133 sgen_find_section_pin_queue_start_end (nursery_section);
2137 reset_heap_boundaries ();
2138 sgen_update_heap_boundaries ((mword)sgen_get_nursery_start (), (mword)sgen_get_nursery_end ());
2140 /* walk the pin_queue, build up the fragment list of free memory, unmark
2141 * pinned objects as we go, memzero() the empty fragments so they are ready for the
2144 fragment_total = sgen_build_nursery_fragments (nursery_section, NULL);
2145 if (!fragment_total)
2147 SGEN_LOG (4, "Free space in nursery after major %ld", (long)fragment_total);
2149 if (do_concurrent_checks && concurrent_collection_in_progress)
2150 sgen_debug_check_nursery_is_clean ();
2152 /* prepare the pin queue for the next collection */
2153 sgen_finish_pinning ();
2155 /* Clear TLABs for all threads */
2156 sgen_clear_tlabs ();
2158 sgen_pin_stats_reset ();
2160 sgen_cement_clear_below_threshold ();
2162 if (check_mark_bits_after_major_collection)
2163 sgen_check_heap_marked (concurrent_collection_in_progress);
2166 time_major_fragment_creation += TV_ELAPSED (atv, btv);
2168 binary_protocol_sweep_begin (GENERATION_OLD, !major_collector.sweeps_lazily);
2169 sgen_memgov_major_pre_sweep ();
2172 time_major_free_bigobjs += TV_ELAPSED (btv, atv);
2177 time_major_los_sweep += TV_ELAPSED (atv, btv);
2179 major_collector.sweep ();
2181 binary_protocol_sweep_end (GENERATION_OLD, !major_collector.sweeps_lazily);
2184 time_major_sweep += TV_ELAPSED (btv, atv);
2186 sgen_debug_dump_heap ("major", gc_stats.major_gc_count - 1, reason);
2188 if (sgen_have_pending_finalizers ()) {
2189 SGEN_LOG (4, "Finalizer-thread wakeup");
2190 sgen_client_finalize_notify ();
2193 sgen_memgov_major_collection_end (forced, concurrent_collection_in_progress, reason, is_overflow);
2194 current_collection_generation = -1;
2196 memset (&counts, 0, sizeof (ScannedObjectCounts));
2197 major_collector.finish_major_collection (&counts);
2199 sgen_workers_assert_gray_queue_is_empty ();
2201 SGEN_ASSERT (0, sgen_workers_all_done (), "Can't have workers working after major collection has finished");
2202 if (concurrent_collection_in_progress)
2203 concurrent_collection_in_progress = FALSE;
2205 check_scan_starts ();
2207 binary_protocol_flush_buffers (FALSE);
2209 //consistency_check ();
2211 binary_protocol_collection_end (gc_stats.major_gc_count - 1, GENERATION_OLD, counts.num_scanned_objects, counts.num_unique_scanned_objects);
2215 major_do_collection (const char *reason, gboolean is_overflow, gboolean forced)
2217 TV_DECLARE (time_start);
2218 TV_DECLARE (time_end);
2219 size_t old_next_pin_slot;
2220 SgenGrayQueue gc_thread_gray_queue;
2222 if (disable_major_collections)
2225 if (major_collector.get_and_reset_num_major_objects_marked) {
2226 long long num_marked = major_collector.get_and_reset_num_major_objects_marked ();
2227 g_assert (!num_marked);
2230 /* world must be stopped already */
2231 TV_GETTIME (time_start);
2233 init_gray_queue (&gc_thread_gray_queue, FALSE);
2234 major_start_collection (&gc_thread_gray_queue, reason, FALSE, &old_next_pin_slot);
2235 major_finish_collection (&gc_thread_gray_queue, reason, is_overflow, old_next_pin_slot, forced);
2236 sgen_gray_object_queue_dispose (&gc_thread_gray_queue);
2238 TV_GETTIME (time_end);
2239 gc_stats.major_gc_time += TV_ELAPSED (time_start, time_end);
2241 /* FIXME: also report this to the user, preferably in gc-end. */
2242 if (major_collector.get_and_reset_num_major_objects_marked)
2243 major_collector.get_and_reset_num_major_objects_marked ();
2245 return bytes_pinned_from_failed_allocation > 0;
2249 major_start_concurrent_collection (const char *reason)
2251 TV_DECLARE (time_start);
2252 TV_DECLARE (time_end);
2253 long long num_objects_marked;
2254 SgenGrayQueue gc_thread_gray_queue;
2256 if (disable_major_collections)
2259 TV_GETTIME (time_start);
2260 SGEN_TV_GETTIME (time_major_conc_collection_start);
2262 num_objects_marked = major_collector.get_and_reset_num_major_objects_marked ();
2263 g_assert (num_objects_marked == 0);
2265 binary_protocol_concurrent_start ();
2267 init_gray_queue (&gc_thread_gray_queue, TRUE);
2268 // FIXME: store reason and pass it when finishing
2269 major_start_collection (&gc_thread_gray_queue, reason, TRUE, NULL);
2270 sgen_gray_object_queue_dispose (&gc_thread_gray_queue);
2272 num_objects_marked = major_collector.get_and_reset_num_major_objects_marked ();
2274 TV_GETTIME (time_end);
2275 gc_stats.major_gc_time += TV_ELAPSED (time_start, time_end);
2277 current_collection_generation = -1;
2281 * Returns whether the major collection has finished.
2284 major_should_finish_concurrent_collection (void)
2286 return sgen_workers_all_done ();
2290 major_update_concurrent_collection (void)
2292 TV_DECLARE (total_start);
2293 TV_DECLARE (total_end);
2295 TV_GETTIME (total_start);
2297 binary_protocol_concurrent_update ();
2299 major_collector.update_cardtable_mod_union ();
2300 sgen_los_update_cardtable_mod_union ();
2302 TV_GETTIME (total_end);
2303 gc_stats.major_gc_time += TV_ELAPSED (total_start, total_end);
2307 major_finish_concurrent_collection (gboolean forced)
2309 SgenGrayQueue gc_thread_gray_queue;
2310 TV_DECLARE (total_start);
2311 TV_DECLARE (total_end);
2313 TV_GETTIME (total_start);
2315 binary_protocol_concurrent_finish ();
2318 * We need to stop all workers since we're updating the cardtable below.
2319 * The workers will be resumed with a finishing pause context to avoid
2320 * additional cardtable and object scanning.
2322 sgen_workers_stop_all_workers ();
2324 SGEN_TV_GETTIME (time_major_conc_collection_end);
2325 gc_stats.major_gc_time_concurrent += SGEN_TV_ELAPSED (time_major_conc_collection_start, time_major_conc_collection_end);
2327 major_collector.update_cardtable_mod_union ();
2328 sgen_los_update_cardtable_mod_union ();
2330 if (mod_union_consistency_check)
2331 sgen_check_mod_union_consistency ();
2333 current_collection_generation = GENERATION_OLD;
2334 sgen_cement_reset ();
2335 init_gray_queue (&gc_thread_gray_queue, FALSE);
2336 major_finish_collection (&gc_thread_gray_queue, "finishing", FALSE, -1, forced);
2337 sgen_gray_object_queue_dispose (&gc_thread_gray_queue);
2339 TV_GETTIME (total_end);
2340 gc_stats.major_gc_time += TV_ELAPSED (total_start, total_end);
2342 current_collection_generation = -1;
2346 * Ensure an allocation request for @size will succeed by freeing enough memory.
2348 * LOCKING: The GC lock MUST be held.
2351 sgen_ensure_free_space (size_t size, int generation)
2353 int generation_to_collect = -1;
2354 const char *reason = NULL;
2356 if (generation == GENERATION_OLD) {
2357 if (sgen_need_major_collection (size)) {
2358 reason = "LOS overflow";
2359 generation_to_collect = GENERATION_OLD;
2362 if (degraded_mode) {
2363 if (sgen_need_major_collection (size)) {
2364 reason = "Degraded mode overflow";
2365 generation_to_collect = GENERATION_OLD;
2367 } else if (sgen_need_major_collection (size)) {
2368 reason = concurrent_collection_in_progress ? "Forced finish concurrent collection" : "Minor allowance";
2369 generation_to_collect = GENERATION_OLD;
2371 generation_to_collect = GENERATION_NURSERY;
2372 reason = "Nursery full";
2376 if (generation_to_collect == -1) {
2377 if (concurrent_collection_in_progress && sgen_workers_all_done ()) {
2378 generation_to_collect = GENERATION_OLD;
2379 reason = "Finish concurrent collection";
2383 if (generation_to_collect == -1)
2385 sgen_perform_collection (size, generation_to_collect, reason, FALSE, TRUE);
2389 * LOCKING: Assumes the GC lock is held.
2392 sgen_perform_collection (size_t requested_size, int generation_to_collect, const char *reason, gboolean wait_to_finish, gboolean stw)
2394 TV_DECLARE (gc_total_start);
2395 TV_DECLARE (gc_total_end);
2396 int overflow_generation_to_collect = -1;
2397 int oldest_generation_collected = generation_to_collect;
2398 const char *overflow_reason = NULL;
2399 gboolean finish_concurrent = concurrent_collection_in_progress && (major_should_finish_concurrent_collection () || generation_to_collect == GENERATION_OLD);
2401 binary_protocol_collection_requested (generation_to_collect, requested_size, wait_to_finish ? 1 : 0);
2403 SGEN_ASSERT (0, generation_to_collect == GENERATION_NURSERY || generation_to_collect == GENERATION_OLD, "What generation is this?");
2406 sgen_stop_world (generation_to_collect);
2408 SGEN_ASSERT (0, sgen_is_world_stopped (), "We can only collect if the world is stopped");
2411 TV_GETTIME (gc_total_start);
2413 // FIXME: extract overflow reason
2414 // FIXME: minor overflow for concurrent case
2415 if (generation_to_collect == GENERATION_NURSERY && !finish_concurrent) {
2416 if (concurrent_collection_in_progress)
2417 major_update_concurrent_collection ();
2419 if (collect_nursery (reason, FALSE, NULL) && !concurrent_collection_in_progress) {
2420 overflow_generation_to_collect = GENERATION_OLD;
2421 overflow_reason = "Minor overflow";
2423 } else if (finish_concurrent) {
2424 major_finish_concurrent_collection (wait_to_finish);
2425 oldest_generation_collected = GENERATION_OLD;
2427 SGEN_ASSERT (0, generation_to_collect == GENERATION_OLD, "We should have handled nursery collections above");
2428 if (major_collector.is_concurrent && !wait_to_finish) {
2429 collect_nursery ("Concurrent start", FALSE, NULL);
2430 major_start_concurrent_collection (reason);
2431 oldest_generation_collected = GENERATION_NURSERY;
2432 } else if (major_do_collection (reason, FALSE, wait_to_finish)) {
2433 overflow_generation_to_collect = GENERATION_NURSERY;
2434 overflow_reason = "Excessive pinning";
2438 if (overflow_generation_to_collect != -1) {
2439 SGEN_ASSERT (0, !concurrent_collection_in_progress, "We don't yet support overflow collections with the concurrent collector");
2442 * We need to do an overflow collection, either because we ran out of memory
2443 * or the nursery is fully pinned.
2446 if (overflow_generation_to_collect == GENERATION_NURSERY)
2447 collect_nursery (overflow_reason, TRUE, NULL);
2449 major_do_collection (overflow_reason, TRUE, wait_to_finish);
2451 oldest_generation_collected = MAX (oldest_generation_collected, overflow_generation_to_collect);
2454 SGEN_LOG (2, "Heap size: %lu, LOS size: %lu", (unsigned long)sgen_gc_get_total_heap_allocation (), (unsigned long)los_memory_usage);
2456 /* this also sets the proper pointers for the next allocation */
2457 if (generation_to_collect == GENERATION_NURSERY && !sgen_can_alloc_size (requested_size)) {
2458 /* TypeBuilder and MonoMethod are killing mcs with fragmentation */
2459 SGEN_LOG (1, "nursery collection didn't find enough room for %zd alloc (%zd pinned)", requested_size, sgen_get_pinned_count ());
2460 sgen_dump_pin_queue ();
2464 TV_GETTIME (gc_total_end);
2465 time_max = MAX (time_max, TV_ELAPSED (gc_total_start, gc_total_end));
2468 sgen_restart_world (oldest_generation_collected);
2472 * ######################################################################
2473 * ######## Memory allocation from the OS
2474 * ######################################################################
2475 * This section of code deals with getting memory from the OS and
2476 * allocating memory for GC-internal data structures.
2477 * Internal memory can be handled with a freelist for small objects.
2483 G_GNUC_UNUSED static void
2484 report_internal_mem_usage (void)
2486 printf ("Internal memory usage:\n");
2487 sgen_report_internal_mem_usage ();
2488 printf ("Pinned memory usage:\n");
2489 major_collector.report_pinned_memory_usage ();
2493 * ######################################################################
2494 * ######## Finalization support
2495 * ######################################################################
2499 * If the object has been forwarded it means it's still referenced from a root.
2500 * If it is pinned it's still alive as well.
2501 * A LOS object is only alive if we have pinned it.
2502 * Return TRUE if @obj is ready to be finalized.
2504 static inline gboolean
2505 sgen_is_object_alive (GCObject *object)
2507 if (ptr_in_nursery (object))
2508 return sgen_nursery_is_object_alive (object);
2510 return sgen_major_is_object_alive (object);
2514 * This function returns true if @object is either alive and belongs to the
2515 * current collection - major collections are full heap, so old gen objects
2516 * are never alive during a minor collection.
2519 sgen_is_object_alive_and_on_current_collection (GCObject *object)
2521 if (ptr_in_nursery (object))
2522 return sgen_nursery_is_object_alive (object);
2524 if (current_collection_generation == GENERATION_NURSERY)
2527 return sgen_major_is_object_alive (object);
2532 sgen_gc_is_object_ready_for_finalization (GCObject *object)
2534 return !sgen_is_object_alive (object);
2538 sgen_queue_finalization_entry (GCObject *obj)
2540 gboolean critical = sgen_client_object_has_critical_finalizer (obj);
2542 sgen_pointer_queue_add (critical ? &critical_fin_queue : &fin_ready_queue, obj);
2544 sgen_client_object_queued_for_finalization (obj);
2548 sgen_object_is_live (GCObject *obj)
2550 return sgen_is_object_alive_and_on_current_collection (obj);
2554 * `System.GC.WaitForPendingFinalizers` first checks `sgen_have_pending_finalizers()` to
2555 * determine whether it can exit quickly. The latter must therefore only return FALSE if
2556 * all finalizers have really finished running.
2558 * `sgen_gc_invoke_finalizers()` first dequeues a finalizable object, and then finalizes it.
2559 * This means that just checking whether the queues are empty leaves the possibility that an
2560 * object might have been dequeued but not yet finalized. That's why we need the additional
2561 * flag `pending_unqueued_finalizer`.
2564 static volatile gboolean pending_unqueued_finalizer = FALSE;
2565 volatile gboolean sgen_suspend_finalizers = FALSE;
2568 sgen_set_suspend_finalizers (void)
2570 sgen_suspend_finalizers = TRUE;
2574 sgen_gc_invoke_finalizers (void)
2578 g_assert (!pending_unqueued_finalizer);
2580 /* FIXME: batch to reduce lock contention */
2581 while (sgen_have_pending_finalizers ()) {
2587 * We need to set `pending_unqueued_finalizer` before dequeing the
2588 * finalizable object.
2590 if (!sgen_pointer_queue_is_empty (&fin_ready_queue)) {
2591 pending_unqueued_finalizer = TRUE;
2592 mono_memory_write_barrier ();
2593 obj = (GCObject *)sgen_pointer_queue_pop (&fin_ready_queue);
2594 } else if (!sgen_pointer_queue_is_empty (&critical_fin_queue)) {
2595 pending_unqueued_finalizer = TRUE;
2596 mono_memory_write_barrier ();
2597 obj = (GCObject *)sgen_pointer_queue_pop (&critical_fin_queue);
2603 SGEN_LOG (7, "Finalizing object %p (%s)", obj, sgen_client_vtable_get_name (SGEN_LOAD_VTABLE (obj)));
2611 /* the object is on the stack so it is pinned */
2612 /*g_print ("Calling finalizer for object: %p (%s)\n", obj, sgen_client_object_safe_name (obj));*/
2613 sgen_client_run_finalize (obj);
2616 if (pending_unqueued_finalizer) {
2617 mono_memory_write_barrier ();
2618 pending_unqueued_finalizer = FALSE;
2625 sgen_have_pending_finalizers (void)
2627 if (sgen_suspend_finalizers)
2629 return pending_unqueued_finalizer || !sgen_pointer_queue_is_empty (&fin_ready_queue) || !sgen_pointer_queue_is_empty (&critical_fin_queue);
2633 * ######################################################################
2634 * ######## registered roots support
2635 * ######################################################################
2639 * We do not coalesce roots.
2642 sgen_register_root (char *start, size_t size, SgenDescriptor descr, int root_type, int source, const char *msg)
2644 RootRecord new_root;
2647 for (i = 0; i < ROOT_TYPE_NUM; ++i) {
2648 RootRecord *root = (RootRecord *)sgen_hash_table_lookup (&roots_hash [i], start);
2649 /* we allow changing the size and the descriptor (for thread statics etc) */
2651 size_t old_size = root->end_root - start;
2652 root->end_root = start + size;
2653 SGEN_ASSERT (0, !!root->root_desc == !!descr, "Can't change whether a root is precise or conservative.");
2654 SGEN_ASSERT (0, root->source == source, "Can't change a root's source identifier.");
2655 SGEN_ASSERT (0, !!root->msg == !!msg, "Can't change a root's message.");
2656 root->root_desc = descr;
2658 roots_size -= old_size;
2664 new_root.end_root = start + size;
2665 new_root.root_desc = descr;
2666 new_root.source = source;
2669 sgen_hash_table_replace (&roots_hash [root_type], start, &new_root, NULL);
2672 SGEN_LOG (3, "Added root for range: %p-%p, descr: %llx (%d/%d bytes)", start, new_root.end_root, (long long)descr, (int)size, (int)roots_size);
2679 sgen_deregister_root (char* addr)
2685 for (root_type = 0; root_type < ROOT_TYPE_NUM; ++root_type) {
2686 if (sgen_hash_table_remove (&roots_hash [root_type], addr, &root))
2687 roots_size -= (root.end_root - addr);
2693 sgen_wbroots_iterate_live_block_ranges (sgen_cardtable_block_callback cb)
2697 SGEN_HASH_TABLE_FOREACH (&roots_hash [ROOT_TYPE_WBARRIER], void **, start_root, RootRecord *, root) {
2698 cb ((mword)start_root, (mword)root->end_root - (mword)start_root);
2699 } SGEN_HASH_TABLE_FOREACH_END;
2702 /* Root equivalent of sgen_client_cardtable_scan_object */
2704 sgen_wbroot_scan_card_table (void** start_root, mword size, ScanCopyContext ctx)
2706 ScanPtrFieldFunc scan_field_func = ctx.ops->scan_ptr_field;
2707 guint8 *card_data = sgen_card_table_get_card_scan_address ((mword)start_root);
2708 guint8 *card_base = card_data;
2709 mword card_count = sgen_card_table_number_of_cards_in_range ((mword)start_root, size);
2710 guint8 *card_data_end = card_data + card_count;
2711 mword extra_idx = 0;
2712 char *obj_start = sgen_card_table_align_pointer (start_root);
2713 char *obj_end = (char*)start_root + size;
2714 #ifdef SGEN_HAVE_OVERLAPPING_CARDS
2715 guint8 *overflow_scan_end = NULL;
2718 #ifdef SGEN_HAVE_OVERLAPPING_CARDS
2719 /*Check for overflow and if so, setup to scan in two steps*/
2720 if (card_data_end >= SGEN_SHADOW_CARDTABLE_END) {
2721 overflow_scan_end = sgen_shadow_cardtable + (card_data_end - SGEN_SHADOW_CARDTABLE_END);
2722 card_data_end = SGEN_SHADOW_CARDTABLE_END;
2728 card_data = sgen_find_next_card (card_data, card_data_end);
2730 for (; card_data < card_data_end; card_data = sgen_find_next_card (card_data + 1, card_data_end)) {
2731 size_t idx = (card_data - card_base) + extra_idx;
2732 char *start = (char*)(obj_start + idx * CARD_SIZE_IN_BYTES);
2733 char *card_end = start + CARD_SIZE_IN_BYTES;
2734 char *elem = start, *first_elem = start;
2737 * Don't clean first and last card on 32bit systems since they
2738 * may also be part from other roots.
2740 if (card_data != card_base && card_data != (card_data_end - 1))
2741 sgen_card_table_prepare_card_for_scanning (card_data);
2743 card_end = MIN (card_end, obj_end);
2745 if (elem < (char*)start_root)
2746 first_elem = elem = (char*)start_root;
2748 for (; elem < card_end; elem += SIZEOF_VOID_P) {
2749 if (*(GCObject**)elem)
2750 scan_field_func (NULL, (GCObject**)elem, ctx.queue);
2753 binary_protocol_card_scan (first_elem, elem - first_elem);
2756 #ifdef SGEN_HAVE_OVERLAPPING_CARDS
2757 if (overflow_scan_end) {
2758 extra_idx = card_data - card_base;
2759 card_base = card_data = sgen_shadow_cardtable;
2760 card_data_end = overflow_scan_end;
2761 overflow_scan_end = NULL;
2768 sgen_wbroots_scan_card_table (ScanCopyContext ctx)
2773 SGEN_HASH_TABLE_FOREACH (&roots_hash [ROOT_TYPE_WBARRIER], void **, start_root, RootRecord *, root) {
2774 SGEN_ASSERT (0, (root->root_desc & ROOT_DESC_TYPE_MASK) == ROOT_DESC_VECTOR, "Unsupported root type");
2776 sgen_wbroot_scan_card_table (start_root, (mword)root->end_root - (mword)start_root, ctx);
2777 } SGEN_HASH_TABLE_FOREACH_END;
2781 * ######################################################################
2782 * ######## Thread handling (stop/start code)
2783 * ######################################################################
2787 sgen_get_current_collection_generation (void)
2789 return current_collection_generation;
2793 sgen_thread_register (SgenThreadInfo* info, void *stack_bottom_fallback)
2795 info->tlab_start = info->tlab_next = info->tlab_temp_end = info->tlab_real_end = NULL;
2797 sgen_client_thread_register (info, stack_bottom_fallback);
2803 sgen_thread_unregister (SgenThreadInfo *p)
2805 sgen_client_thread_unregister (p);
2809 * ######################################################################
2810 * ######## Write barriers
2811 * ######################################################################
2815 * Note: the write barriers first do the needed GC work and then do the actual store:
2816 * this way the value is visible to the conservative GC scan after the write barrier
2817 * itself. If a GC interrupts the barrier in the middle, value will be kept alive by
2818 * the conservative scan, otherwise by the remembered set scan.
2822 * mono_gc_wbarrier_arrayref_copy:
2825 mono_gc_wbarrier_arrayref_copy (gpointer dest_ptr, gpointer src_ptr, int count)
2827 HEAVY_STAT (++stat_wbarrier_arrayref_copy);
2828 /*This check can be done without taking a lock since dest_ptr array is pinned*/
2829 if (ptr_in_nursery (dest_ptr) || count <= 0) {
2830 mono_gc_memmove_aligned (dest_ptr, src_ptr, count * sizeof (gpointer));
2834 #ifdef SGEN_HEAVY_BINARY_PROTOCOL
2835 if (binary_protocol_is_heavy_enabled ()) {
2837 for (i = 0; i < count; ++i) {
2838 gpointer dest = (gpointer*)dest_ptr + i;
2839 gpointer obj = *((gpointer*)src_ptr + i);
2841 binary_protocol_wbarrier (dest, obj, (gpointer)LOAD_VTABLE (obj));
2846 remset.wbarrier_arrayref_copy (dest_ptr, src_ptr, count);
2850 * mono_gc_wbarrier_generic_nostore:
2853 mono_gc_wbarrier_generic_nostore (gpointer ptr)
2857 HEAVY_STAT (++stat_wbarrier_generic_store);
2859 sgen_client_wbarrier_generic_nostore_check (ptr);
2861 obj = *(gpointer*)ptr;
2863 binary_protocol_wbarrier (ptr, obj, (gpointer)LOAD_VTABLE (obj));
2866 * We need to record old->old pointer locations for the
2867 * concurrent collector.
2869 if (!ptr_in_nursery (obj) && !concurrent_collection_in_progress) {
2870 SGEN_LOG (8, "Skipping remset at %p", ptr);
2874 SGEN_LOG (8, "Adding remset at %p", ptr);
2876 remset.wbarrier_generic_nostore (ptr);
2880 * mono_gc_wbarrier_generic_store:
2883 mono_gc_wbarrier_generic_store (gpointer ptr, GCObject* value)
2885 SGEN_LOG (8, "Wbarrier store at %p to %p (%s)", ptr, value, value ? sgen_client_vtable_get_name (SGEN_LOAD_VTABLE (value)) : "null");
2886 SGEN_UPDATE_REFERENCE_ALLOW_NULL (ptr, value);
2887 if (ptr_in_nursery (value) || concurrent_collection_in_progress)
2888 mono_gc_wbarrier_generic_nostore (ptr);
2889 sgen_dummy_use (value);
2893 * mono_gc_wbarrier_generic_store_atomic:
2894 * Same as \c mono_gc_wbarrier_generic_store but performs the store
2895 * as an atomic operation with release semantics.
2898 mono_gc_wbarrier_generic_store_atomic (gpointer ptr, GCObject *value)
2900 HEAVY_STAT (++stat_wbarrier_generic_store_atomic);
2902 SGEN_LOG (8, "Wbarrier atomic store at %p to %p (%s)", ptr, value, value ? sgen_client_vtable_get_name (SGEN_LOAD_VTABLE (value)) : "null");
2904 InterlockedWritePointer ((volatile gpointer *)ptr, value);
2906 if (ptr_in_nursery (value) || concurrent_collection_in_progress)
2907 mono_gc_wbarrier_generic_nostore (ptr);
2909 sgen_dummy_use (value);
2913 sgen_wbarrier_value_copy_bitmap (gpointer _dest, gpointer _src, int size, unsigned bitmap)
2915 GCObject **dest = (GCObject **)_dest;
2916 GCObject **src = (GCObject **)_src;
2920 mono_gc_wbarrier_generic_store (dest, *src);
2925 size -= SIZEOF_VOID_P;
2931 * ######################################################################
2932 * ######## Other mono public interface functions.
2933 * ######################################################################
2937 sgen_gc_collect (int generation)
2942 sgen_perform_collection (0, generation, "user request", TRUE, TRUE);
2947 sgen_gc_collection_count (int generation)
2949 if (generation == 0)
2950 return gc_stats.minor_gc_count;
2951 return gc_stats.major_gc_count;
2955 sgen_gc_get_used_size (void)
2959 tot = los_memory_usage;
2960 tot += nursery_section->next_data - nursery_section->data;
2961 tot += major_collector.get_used_size ();
2962 /* FIXME: account for pinned objects */
2968 sgen_env_var_error (const char *env_var, const char *fallback, const char *description_format, ...)
2972 va_start (ap, description_format);
2974 fprintf (stderr, "Warning: In environment variable `%s': ", env_var);
2975 vfprintf (stderr, description_format, ap);
2977 fprintf (stderr, " - %s", fallback);
2978 fprintf (stderr, "\n");
2984 parse_double_in_interval (const char *env_var, const char *opt_name, const char *opt, double min, double max, double *result)
2987 double val = strtod (opt, &endptr);
2988 if (endptr == opt) {
2989 sgen_env_var_error (env_var, "Using default value.", "`%s` must be a number.", opt_name);
2992 else if (val < min || val > max) {
2993 sgen_env_var_error (env_var, "Using default value.", "`%s` must be between %.2f - %.2f.", opt_name, min, max);
3005 char *major_collector_opt = NULL;
3006 char *minor_collector_opt = NULL;
3007 char *params_opts = NULL;
3008 char *debug_opts = NULL;
3009 size_t max_heap = 0;
3010 size_t soft_limit = 0;
3012 gboolean debug_print_allowance = FALSE;
3013 double allowance_ratio = 0, save_target = 0;
3014 gboolean cement_enabled = TRUE;
3017 result = InterlockedCompareExchange (&gc_initialized, -1, 0);
3020 /* already inited */
3023 /* being inited by another thread */
3024 mono_thread_info_usleep (1000);
3027 /* we will init it */
3030 g_assert_not_reached ();
3032 } while (result != 0);
3034 SGEN_TV_GETTIME (sgen_init_timestamp);
3036 #ifdef SGEN_WITHOUT_MONO
3037 mono_thread_smr_init ();
3040 mono_coop_mutex_init (&gc_mutex);
3042 gc_debug_file = stderr;
3044 mono_coop_mutex_init (&sgen_interruption_mutex);
3046 if ((env = g_getenv (MONO_GC_PARAMS_NAME)) || gc_params_options) {
3047 params_opts = g_strdup_printf ("%s,%s", gc_params_options ? gc_params_options : "", env ? env : "");
3052 opts = g_strsplit (params_opts, ",", -1);
3053 for (ptr = opts; *ptr; ++ptr) {
3055 if (g_str_has_prefix (opt, "major=")) {
3056 opt = strchr (opt, '=') + 1;
3057 major_collector_opt = g_strdup (opt);
3058 } else if (g_str_has_prefix (opt, "minor=")) {
3059 opt = strchr (opt, '=') + 1;
3060 minor_collector_opt = g_strdup (opt);
3068 sgen_init_internal_allocator ();
3069 sgen_init_nursery_allocator ();
3070 sgen_init_fin_weak_hash ();
3071 sgen_init_hash_table ();
3072 sgen_init_descriptors ();
3073 sgen_init_gray_queues ();
3074 sgen_init_allocator ();
3075 sgen_init_gchandles ();
3077 sgen_register_fixed_internal_mem_type (INTERNAL_MEM_SECTION, SGEN_SIZEOF_GC_MEM_SECTION);
3078 sgen_register_fixed_internal_mem_type (INTERNAL_MEM_GRAY_QUEUE, sizeof (GrayQueueSection));
3080 sgen_client_init ();
3082 if (!minor_collector_opt) {
3083 sgen_simple_nursery_init (&sgen_minor_collector, FALSE);
3085 if (!strcmp (minor_collector_opt, "simple")) {
3087 sgen_simple_nursery_init (&sgen_minor_collector, FALSE);
3088 } else if (!strcmp (minor_collector_opt, "simple-par")) {
3089 sgen_simple_nursery_init (&sgen_minor_collector, TRUE);
3090 } else if (!strcmp (minor_collector_opt, "split")) {
3091 sgen_split_nursery_init (&sgen_minor_collector);
3093 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Using `simple` instead.", "Unknown minor collector `%s'.", minor_collector_opt);
3094 goto use_simple_nursery;
3098 if (!major_collector_opt) {
3100 DEFAULT_MAJOR_INIT (&major_collector);
3101 } else if (!strcmp (major_collector_opt, "marksweep")) {
3102 sgen_marksweep_init (&major_collector);
3103 } else if (!strcmp (major_collector_opt, "marksweep-conc")) {
3104 sgen_marksweep_conc_init (&major_collector);
3105 } else if (!strcmp (major_collector_opt, "marksweep-conc-par")) {
3106 sgen_marksweep_conc_par_init (&major_collector);
3108 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Using `" DEFAULT_MAJOR_NAME "` instead.", "Unknown major collector `%s'.", major_collector_opt);
3109 goto use_default_major;
3112 sgen_nursery_size = DEFAULT_NURSERY_SIZE;
3115 gboolean usage_printed = FALSE;
3117 for (ptr = opts; *ptr; ++ptr) {
3119 if (!strcmp (opt, ""))
3121 if (g_str_has_prefix (opt, "major="))
3123 if (g_str_has_prefix (opt, "minor="))
3125 if (g_str_has_prefix (opt, "max-heap-size=")) {
3126 size_t page_size = mono_pagesize ();
3127 size_t max_heap_candidate = 0;
3128 opt = strchr (opt, '=') + 1;
3129 if (*opt && mono_gc_parse_environment_string_extract_number (opt, &max_heap_candidate)) {
3130 max_heap = (max_heap_candidate + page_size - 1) & ~(size_t)(page_size - 1);
3131 if (max_heap != max_heap_candidate)
3132 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Rounding up.", "`max-heap-size` size must be a multiple of %d.", page_size);
3134 sgen_env_var_error (MONO_GC_PARAMS_NAME, NULL, "`max-heap-size` must be an integer.");
3138 if (g_str_has_prefix (opt, "soft-heap-limit=")) {
3139 opt = strchr (opt, '=') + 1;
3140 if (*opt && mono_gc_parse_environment_string_extract_number (opt, &soft_limit)) {
3141 if (soft_limit <= 0) {
3142 sgen_env_var_error (MONO_GC_PARAMS_NAME, NULL, "`soft-heap-limit` must be positive.");
3146 sgen_env_var_error (MONO_GC_PARAMS_NAME, NULL, "`soft-heap-limit` must be an integer.");
3152 if (g_str_has_prefix (opt, "nursery-size=")) {
3154 opt = strchr (opt, '=') + 1;
3155 if (*opt && mono_gc_parse_environment_string_extract_number (opt, &val)) {
3156 if ((val & (val - 1))) {
3157 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Using default value.", "`nursery-size` must be a power of two.");
3161 if (val < SGEN_MAX_NURSERY_WASTE) {
3162 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Using default value.",
3163 "`nursery-size` must be at least %d bytes.", SGEN_MAX_NURSERY_WASTE);
3167 sgen_nursery_size = val;
3168 sgen_nursery_bits = 0;
3169 while (ONE_P << (++ sgen_nursery_bits) != sgen_nursery_size)
3172 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Using default value.", "`nursery-size` must be an integer.");
3178 if (g_str_has_prefix (opt, "save-target-ratio=")) {
3180 opt = strchr (opt, '=') + 1;
3181 if (parse_double_in_interval (MONO_GC_PARAMS_NAME, "save-target-ratio", opt,
3182 SGEN_MIN_SAVE_TARGET_RATIO, SGEN_MAX_SAVE_TARGET_RATIO, &val)) {
3187 if (g_str_has_prefix (opt, "default-allowance-ratio=")) {
3189 opt = strchr (opt, '=') + 1;
3190 if (parse_double_in_interval (MONO_GC_PARAMS_NAME, "default-allowance-ratio", opt,
3191 SGEN_MIN_ALLOWANCE_NURSERY_SIZE_RATIO, SGEN_MAX_ALLOWANCE_NURSERY_SIZE_RATIO, &val)) {
3192 allowance_ratio = val;
3197 if (!strcmp (opt, "cementing")) {
3198 cement_enabled = TRUE;
3201 if (!strcmp (opt, "no-cementing")) {
3202 cement_enabled = FALSE;
3206 if (!strcmp (opt, "precleaning")) {
3207 precleaning_enabled = TRUE;
3210 if (!strcmp (opt, "no-precleaning")) {
3211 precleaning_enabled = FALSE;
3215 if (major_collector.handle_gc_param && major_collector.handle_gc_param (opt))
3218 if (sgen_minor_collector.handle_gc_param && sgen_minor_collector.handle_gc_param (opt))
3221 if (sgen_client_handle_gc_param (opt))
3224 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Ignoring.", "Unknown option `%s`.", opt);
3229 fprintf (stderr, "\n%s must be a comma-delimited list of one or more of the following:\n", MONO_GC_PARAMS_NAME);
3230 fprintf (stderr, " max-heap-size=N (where N is an integer, possibly with a k, m or a g suffix)\n");
3231 fprintf (stderr, " soft-heap-limit=n (where N is an integer, possibly with a k, m or a g suffix)\n");
3232 fprintf (stderr, " nursery-size=N (where N is an integer, possibly with a k, m or a g suffix)\n");
3233 fprintf (stderr, " major=COLLECTOR (where COLLECTOR is `marksweep', `marksweep-conc', `marksweep-par')\n");
3234 fprintf (stderr, " minor=COLLECTOR (where COLLECTOR is `simple' or `split')\n");
3235 fprintf (stderr, " wbarrier=WBARRIER (where WBARRIER is `remset' or `cardtable')\n");
3236 fprintf (stderr, " [no-]cementing\n");
3237 if (major_collector.print_gc_param_usage)
3238 major_collector.print_gc_param_usage ();
3239 if (sgen_minor_collector.print_gc_param_usage)
3240 sgen_minor_collector.print_gc_param_usage ();
3241 sgen_client_print_gc_params_usage ();
3242 fprintf (stderr, " Experimental options:\n");
3243 fprintf (stderr, " save-target-ratio=R (where R must be between %.2f - %.2f).\n", SGEN_MIN_SAVE_TARGET_RATIO, SGEN_MAX_SAVE_TARGET_RATIO);
3244 fprintf (stderr, " default-allowance-ratio=R (where R must be between %.2f - %.2f).\n", SGEN_MIN_ALLOWANCE_NURSERY_SIZE_RATIO, SGEN_MAX_ALLOWANCE_NURSERY_SIZE_RATIO);
3245 fprintf (stderr, "\n");
3247 usage_printed = TRUE;
3252 if (major_collector_opt)
3253 g_free (major_collector_opt);
3255 if (minor_collector_opt)
3256 g_free (minor_collector_opt);
3259 g_free (params_opts);
3263 sgen_pinning_init ();
3264 sgen_cement_init (cement_enabled);
3266 if ((env = g_getenv (MONO_GC_DEBUG_NAME)) || gc_debug_options) {
3267 debug_opts = g_strdup_printf ("%s,%s", gc_debug_options ? gc_debug_options : "", env ? env : "");
3272 gboolean usage_printed = FALSE;
3274 opts = g_strsplit (debug_opts, ",", -1);
3275 for (ptr = opts; ptr && *ptr; ptr ++) {
3277 if (!strcmp (opt, ""))
3279 if (opt [0] >= '0' && opt [0] <= '9') {
3280 gc_debug_level = atoi (opt);
3285 char *rf = g_strdup_printf ("%s.%d", opt, mono_process_current_pid ());
3286 gc_debug_file = fopen (rf, "wb");
3288 gc_debug_file = stderr;
3291 } else if (!strcmp (opt, "print-allowance")) {
3292 debug_print_allowance = TRUE;
3293 } else if (!strcmp (opt, "print-pinning")) {
3294 sgen_pin_stats_enable ();
3295 } else if (!strcmp (opt, "verify-before-allocs")) {
3296 verify_before_allocs = 1;
3297 has_per_allocation_action = TRUE;
3298 } else if (g_str_has_prefix (opt, "max-valloc-size=")) {
3299 size_t max_valloc_size;
3300 char *arg = strchr (opt, '=') + 1;
3301 if (*opt && mono_gc_parse_environment_string_extract_number (arg, &max_valloc_size)) {
3302 mono_valloc_set_limit (max_valloc_size);
3304 sgen_env_var_error (MONO_GC_DEBUG_NAME, NULL, "`max-valloc-size` must be an integer.");
3307 } else if (g_str_has_prefix (opt, "verify-before-allocs=")) {
3308 char *arg = strchr (opt, '=') + 1;
3309 verify_before_allocs = atoi (arg);
3310 has_per_allocation_action = TRUE;
3311 } else if (!strcmp (opt, "collect-before-allocs")) {
3312 collect_before_allocs = 1;
3313 has_per_allocation_action = TRUE;
3314 } else if (g_str_has_prefix (opt, "collect-before-allocs=")) {
3315 char *arg = strchr (opt, '=') + 1;
3316 has_per_allocation_action = TRUE;
3317 collect_before_allocs = atoi (arg);
3318 } else if (!strcmp (opt, "verify-before-collections")) {
3319 whole_heap_check_before_collection = TRUE;
3320 } else if (!strcmp (opt, "check-remset-consistency")) {
3321 remset_consistency_checks = TRUE;
3322 nursery_clear_policy = CLEAR_AT_GC;
3323 } else if (!strcmp (opt, "mod-union-consistency-check")) {
3324 if (!major_collector.is_concurrent) {
3325 sgen_env_var_error (MONO_GC_DEBUG_NAME, "Ignoring.", "`mod-union-consistency-check` only works with concurrent major collector.");
3328 mod_union_consistency_check = TRUE;
3329 } else if (!strcmp (opt, "check-mark-bits")) {
3330 check_mark_bits_after_major_collection = TRUE;
3331 } else if (!strcmp (opt, "check-nursery-pinned")) {
3332 check_nursery_objects_pinned = TRUE;
3333 } else if (!strcmp (opt, "clear-at-gc")) {
3334 nursery_clear_policy = CLEAR_AT_GC;
3335 } else if (!strcmp (opt, "clear-nursery-at-gc")) {
3336 nursery_clear_policy = CLEAR_AT_GC;
3337 } else if (!strcmp (opt, "clear-at-tlab-creation")) {
3338 nursery_clear_policy = CLEAR_AT_TLAB_CREATION;
3339 } else if (!strcmp (opt, "debug-clear-at-tlab-creation")) {
3340 nursery_clear_policy = CLEAR_AT_TLAB_CREATION_DEBUG;
3341 } else if (!strcmp (opt, "check-scan-starts")) {
3342 do_scan_starts_check = TRUE;
3343 } else if (!strcmp (opt, "verify-nursery-at-minor-gc")) {
3344 do_verify_nursery = TRUE;
3345 } else if (!strcmp (opt, "check-concurrent")) {
3346 if (!major_collector.is_concurrent) {
3347 sgen_env_var_error (MONO_GC_DEBUG_NAME, "Ignoring.", "`check-concurrent` only works with concurrent major collectors.");
3350 nursery_clear_policy = CLEAR_AT_GC;
3351 do_concurrent_checks = TRUE;
3352 } else if (!strcmp (opt, "dump-nursery-at-minor-gc")) {
3353 do_dump_nursery_content = TRUE;
3354 } else if (!strcmp (opt, "disable-minor")) {
3355 disable_minor_collections = TRUE;
3356 } else if (!strcmp (opt, "disable-major")) {
3357 disable_major_collections = TRUE;
3358 } else if (g_str_has_prefix (opt, "heap-dump=")) {
3359 char *filename = strchr (opt, '=') + 1;
3360 nursery_clear_policy = CLEAR_AT_GC;
3361 sgen_debug_enable_heap_dump (filename);
3362 } else if (g_str_has_prefix (opt, "binary-protocol=")) {
3363 char *filename = strchr (opt, '=') + 1;
3364 char *colon = strrchr (filename, ':');
3367 if (!mono_gc_parse_environment_string_extract_number (colon + 1, &limit)) {
3368 sgen_env_var_error (MONO_GC_DEBUG_NAME, "Ignoring limit.", "Binary protocol file size limit must be an integer.");
3373 binary_protocol_init (filename, (long long)limit);
3374 } else if (!strcmp (opt, "nursery-canaries")) {
3375 do_verify_nursery = TRUE;
3376 enable_nursery_canaries = TRUE;
3377 } else if (!sgen_client_handle_gc_debug (opt)) {
3378 sgen_env_var_error (MONO_GC_DEBUG_NAME, "Ignoring.", "Unknown option `%s`.", opt);
3383 fprintf (stderr, "\n%s must be of the format [<l>[:<filename>]|<option>]+ where <l> is a debug level 0-9.\n", MONO_GC_DEBUG_NAME);
3384 fprintf (stderr, "Valid <option>s are:\n");
3385 fprintf (stderr, " collect-before-allocs[=<n>]\n");
3386 fprintf (stderr, " verify-before-allocs[=<n>]\n");
3387 fprintf (stderr, " max-valloc-size=N (where N is an integer, possibly with a k, m or a g suffix)\n");
3388 fprintf (stderr, " check-remset-consistency\n");
3389 fprintf (stderr, " check-mark-bits\n");
3390 fprintf (stderr, " check-nursery-pinned\n");
3391 fprintf (stderr, " verify-before-collections\n");
3392 fprintf (stderr, " verify-nursery-at-minor-gc\n");
3393 fprintf (stderr, " dump-nursery-at-minor-gc\n");
3394 fprintf (stderr, " disable-minor\n");
3395 fprintf (stderr, " disable-major\n");
3396 fprintf (stderr, " check-concurrent\n");
3397 fprintf (stderr, " clear-[nursery-]at-gc\n");
3398 fprintf (stderr, " clear-at-tlab-creation\n");
3399 fprintf (stderr, " debug-clear-at-tlab-creation\n");
3400 fprintf (stderr, " check-scan-starts\n");
3401 fprintf (stderr, " print-allowance\n");
3402 fprintf (stderr, " print-pinning\n");
3403 fprintf (stderr, " heap-dump=<filename>\n");
3404 fprintf (stderr, " binary-protocol=<filename>[:<file-size-limit>]\n");
3405 fprintf (stderr, " nursery-canaries\n");
3406 sgen_client_print_gc_debug_usage ();
3407 fprintf (stderr, "\n");
3409 usage_printed = TRUE;
3416 g_free (debug_opts);
3418 if (check_mark_bits_after_major_collection)
3419 nursery_clear_policy = CLEAR_AT_GC;
3421 if (major_collector.post_param_init)
3422 major_collector.post_param_init (&major_collector);
3424 if (major_collector.is_concurrent || sgen_minor_collector.is_parallel) {
3425 int num_workers = 1;
3426 if (major_collector.is_parallel || sgen_minor_collector.is_parallel) {
3427 /* FIXME Detect the number of physical cores, instead of logical */
3428 num_workers = mono_cpu_count () / 2;
3429 if (num_workers < 1)
3432 sgen_workers_init (num_workers, (SgenWorkerCallback) major_collector.worker_init_cb);
3435 sgen_memgov_init (max_heap, soft_limit, debug_print_allowance, allowance_ratio, save_target);
3437 memset (&remset, 0, sizeof (remset));
3439 sgen_card_table_init (&remset);
3441 sgen_register_root (NULL, 0, sgen_make_user_root_descriptor (sgen_mark_normal_gc_handles), ROOT_TYPE_NORMAL, MONO_ROOT_SOURCE_GC_HANDLE, "normal gc handles");
3445 sgen_init_bridge ();
3449 sgen_gc_initialized ()
3451 return gc_initialized > 0;
3455 sgen_get_nursery_clear_policy (void)
3457 return nursery_clear_policy;
3463 mono_coop_mutex_lock (&gc_mutex);
3467 sgen_gc_unlock (void)
3469 mono_coop_mutex_unlock (&gc_mutex);
3473 sgen_major_collector_iterate_live_block_ranges (sgen_cardtable_block_callback callback)
3475 major_collector.iterate_live_block_ranges (callback);
3479 sgen_major_collector_iterate_block_ranges (sgen_cardtable_block_callback callback)
3481 major_collector.iterate_block_ranges (callback);
3485 sgen_get_major_collector (void)
3487 return &major_collector;
3491 sgen_get_minor_collector (void)
3493 return &sgen_minor_collector;
3497 sgen_get_remset (void)
3503 count_cards (long long *major_total, long long *major_marked, long long *los_total, long long *los_marked)
3505 sgen_get_major_collector ()->count_cards (major_total, major_marked);
3506 sgen_los_count_cards (los_total, los_marked);
3509 static gboolean world_is_stopped = FALSE;
3511 /* LOCKING: assumes the GC lock is held */
3513 sgen_stop_world (int generation)
3515 long long major_total = -1, major_marked = -1, los_total = -1, los_marked = -1;
3517 SGEN_ASSERT (0, !world_is_stopped, "Why are we stopping a stopped world?");
3519 binary_protocol_world_stopping (generation, sgen_timestamp (), (gpointer) (gsize) mono_native_thread_id_get ());
3521 sgen_client_stop_world (generation);
3523 world_is_stopped = TRUE;
3525 if (binary_protocol_is_heavy_enabled ())
3526 count_cards (&major_total, &major_marked, &los_total, &los_marked);
3527 binary_protocol_world_stopped (generation, sgen_timestamp (), major_total, major_marked, los_total, los_marked);
3530 /* LOCKING: assumes the GC lock is held */
3532 sgen_restart_world (int generation)
3534 long long major_total = -1, major_marked = -1, los_total = -1, los_marked = -1;
3537 SGEN_ASSERT (0, world_is_stopped, "Why are we restarting a running world?");
3539 if (binary_protocol_is_heavy_enabled ())
3540 count_cards (&major_total, &major_marked, &los_total, &los_marked);
3541 binary_protocol_world_restarting (generation, sgen_timestamp (), major_total, major_marked, los_total, los_marked);
3543 world_is_stopped = FALSE;
3545 sgen_client_restart_world (generation, &stw_time);
3547 binary_protocol_world_restarted (generation, sgen_timestamp ());
3549 if (sgen_client_bridge_need_processing ())
3550 sgen_client_bridge_processing_finish (generation);
3552 sgen_memgov_collection_end (generation, stw_time);
3556 sgen_is_world_stopped (void)
3558 return world_is_stopped;
3562 sgen_check_whole_heap_stw (void)
3564 sgen_stop_world (0);
3565 sgen_clear_nursery_fragments ();
3566 sgen_check_whole_heap (TRUE);
3567 sgen_restart_world (0);
3571 sgen_timestamp (void)
3573 SGEN_TV_DECLARE (timestamp);
3574 SGEN_TV_GETTIME (timestamp);
3575 return SGEN_TV_ELAPSED (sgen_init_timestamp, timestamp);
3578 #endif /* HAVE_SGEN_GC */