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>
197 #undef pthread_create
199 #undef pthread_detach
202 * ######################################################################
203 * ######## Types and constants used by the GC.
204 * ######################################################################
207 /* 0 means not initialized, 1 is initialized, -1 means in progress */
208 static int gc_initialized = 0;
209 /* If set, check if we need to do something every X allocations */
210 gboolean has_per_allocation_action;
211 /* If set, do a heap check every X allocation */
212 guint32 verify_before_allocs = 0;
213 /* If set, do a minor collection before every X allocation */
214 guint32 collect_before_allocs = 0;
215 /* If set, do a whole heap check before each collection */
216 static gboolean whole_heap_check_before_collection = FALSE;
217 /* If set, do a remset consistency check at various opportunities */
218 static gboolean remset_consistency_checks = FALSE;
219 /* If set, do a mod union consistency check before each finishing collection pause */
220 static gboolean mod_union_consistency_check = FALSE;
221 /* If set, check whether mark bits are consistent after major collections */
222 static gboolean check_mark_bits_after_major_collection = FALSE;
223 /* If set, check that all nursery objects are pinned/not pinned, depending on context */
224 static gboolean check_nursery_objects_pinned = FALSE;
225 /* If set, do a few checks when the concurrent collector is used */
226 static gboolean do_concurrent_checks = FALSE;
227 /* If set, do a plausibility check on the scan_starts before and after
229 static gboolean do_scan_starts_check = FALSE;
231 static gboolean disable_minor_collections = FALSE;
232 static gboolean disable_major_collections = FALSE;
233 static gboolean do_verify_nursery = FALSE;
234 static gboolean do_dump_nursery_content = FALSE;
235 static gboolean enable_nursery_canaries = FALSE;
237 static gboolean precleaning_enabled = TRUE;
239 #ifdef HEAVY_STATISTICS
240 guint64 stat_objects_alloced_degraded = 0;
241 guint64 stat_bytes_alloced_degraded = 0;
243 guint64 stat_copy_object_called_nursery = 0;
244 guint64 stat_objects_copied_nursery = 0;
245 guint64 stat_copy_object_called_major = 0;
246 guint64 stat_objects_copied_major = 0;
248 guint64 stat_scan_object_called_nursery = 0;
249 guint64 stat_scan_object_called_major = 0;
251 guint64 stat_slots_allocated_in_vain;
253 guint64 stat_nursery_copy_object_failed_from_space = 0;
254 guint64 stat_nursery_copy_object_failed_forwarded = 0;
255 guint64 stat_nursery_copy_object_failed_pinned = 0;
256 guint64 stat_nursery_copy_object_failed_to_space = 0;
258 static guint64 stat_wbarrier_add_to_global_remset = 0;
259 static guint64 stat_wbarrier_arrayref_copy = 0;
260 static guint64 stat_wbarrier_generic_store = 0;
261 static guint64 stat_wbarrier_generic_store_atomic = 0;
262 static guint64 stat_wbarrier_set_root = 0;
265 static guint64 stat_pinned_objects = 0;
267 static guint64 time_minor_pre_collection_fragment_clear = 0;
268 static guint64 time_minor_pinning = 0;
269 static guint64 time_minor_scan_remsets = 0;
270 static guint64 time_minor_scan_pinned = 0;
271 static guint64 time_minor_scan_roots = 0;
272 static guint64 time_minor_finish_gray_stack = 0;
273 static guint64 time_minor_fragment_creation = 0;
275 static guint64 time_major_pre_collection_fragment_clear = 0;
276 static guint64 time_major_pinning = 0;
277 static guint64 time_major_scan_pinned = 0;
278 static guint64 time_major_scan_roots = 0;
279 static guint64 time_major_scan_mod_union = 0;
280 static guint64 time_major_finish_gray_stack = 0;
281 static guint64 time_major_free_bigobjs = 0;
282 static guint64 time_major_los_sweep = 0;
283 static guint64 time_major_sweep = 0;
284 static guint64 time_major_fragment_creation = 0;
286 static guint64 time_max = 0;
288 static SGEN_TV_DECLARE (time_major_conc_collection_start);
289 static SGEN_TV_DECLARE (time_major_conc_collection_end);
291 int gc_debug_level = 0;
293 static char* gc_params_options;
294 static char* gc_debug_options;
298 mono_gc_flush_info (void)
300 fflush (gc_debug_file);
304 #define TV_DECLARE SGEN_TV_DECLARE
305 #define TV_GETTIME SGEN_TV_GETTIME
306 #define TV_ELAPSED SGEN_TV_ELAPSED
308 static SGEN_TV_DECLARE (sgen_init_timestamp);
310 NurseryClearPolicy nursery_clear_policy = CLEAR_AT_TLAB_CREATION;
312 #define object_is_forwarded SGEN_OBJECT_IS_FORWARDED
313 #define object_is_pinned SGEN_OBJECT_IS_PINNED
314 #define pin_object SGEN_PIN_OBJECT
316 #define ptr_in_nursery sgen_ptr_in_nursery
318 #define LOAD_VTABLE SGEN_LOAD_VTABLE
321 nursery_canaries_enabled (void)
323 return enable_nursery_canaries;
326 #define safe_object_get_size sgen_safe_object_get_size
328 #if defined(HAVE_CONC_GC_AS_DEFAULT)
329 /* Use concurrent major on deskstop platforms */
330 #define DEFAULT_MAJOR_INIT sgen_marksweep_conc_init
331 #define DEFAULT_MAJOR_NAME "marksweep-conc"
333 #define DEFAULT_MAJOR_INIT sgen_marksweep_init
334 #define DEFAULT_MAJOR_NAME "marksweep"
338 * ######################################################################
339 * ######## Global data.
340 * ######################################################################
342 MonoCoopMutex gc_mutex;
344 #define SCAN_START_SIZE SGEN_SCAN_START_SIZE
346 size_t degraded_mode = 0;
348 static mword bytes_pinned_from_failed_allocation = 0;
350 GCMemSection *nursery_section = NULL;
351 static volatile mword lowest_heap_address = ~(mword)0;
352 static volatile mword highest_heap_address = 0;
354 MonoCoopMutex sgen_interruption_mutex;
356 int current_collection_generation = -1;
357 static volatile gboolean concurrent_collection_in_progress = FALSE;
359 /* objects that are ready to be finalized */
360 static SgenPointerQueue fin_ready_queue = SGEN_POINTER_QUEUE_INIT (INTERNAL_MEM_FINALIZE_READY);
361 static SgenPointerQueue critical_fin_queue = SGEN_POINTER_QUEUE_INIT (INTERNAL_MEM_FINALIZE_READY);
363 /* registered roots: the key to the hash is the root start address */
365 * Different kinds of roots are kept separate to speed up pin_from_roots () for example.
367 SgenHashTable roots_hash [ROOT_TYPE_NUM] = {
368 SGEN_HASH_TABLE_INIT (INTERNAL_MEM_ROOTS_TABLE, INTERNAL_MEM_ROOT_RECORD, sizeof (RootRecord), sgen_aligned_addr_hash, NULL),
369 SGEN_HASH_TABLE_INIT (INTERNAL_MEM_ROOTS_TABLE, INTERNAL_MEM_ROOT_RECORD, sizeof (RootRecord), sgen_aligned_addr_hash, NULL),
370 SGEN_HASH_TABLE_INIT (INTERNAL_MEM_ROOTS_TABLE, INTERNAL_MEM_ROOT_RECORD, sizeof (RootRecord), sgen_aligned_addr_hash, NULL)
372 static mword roots_size = 0; /* amount of memory in the root set */
374 /* The size of a TLAB */
375 /* The bigger the value, the less often we have to go to the slow path to allocate a new
376 * one, but the more space is wasted by threads not allocating much memory.
378 * FIXME: Make this self-tuning for each thread.
380 guint32 tlab_size = (1024 * 4);
382 #define MAX_SMALL_OBJ_SIZE SGEN_MAX_SMALL_OBJ_SIZE
384 #define ALLOC_ALIGN SGEN_ALLOC_ALIGN
386 #define ALIGN_UP SGEN_ALIGN_UP
388 #ifdef SGEN_DEBUG_INTERNAL_ALLOC
389 MonoNativeThreadId main_gc_thread = NULL;
392 /*Object was pinned during the current collection*/
393 static mword objects_pinned;
396 * ######################################################################
397 * ######## Macros and function declarations.
398 * ######################################################################
401 /* forward declarations */
402 static void scan_from_registered_roots (char *addr_start, char *addr_end, int root_type, ScanCopyContext ctx);
404 static void pin_from_roots (void *start_nursery, void *end_nursery, ScanCopyContext ctx);
405 static void finish_gray_stack (int generation, ScanCopyContext ctx);
408 SgenMajorCollector major_collector;
409 SgenMinorCollector sgen_minor_collector;
411 static SgenRememberedSet remset;
414 * The gray queue a worker job must use. If we're not parallel or
415 * concurrent, we use the main gray queue.
417 static SgenGrayQueue*
418 sgen_workers_get_job_gray_queue (WorkerData *worker_data, SgenGrayQueue *default_gray_queue)
421 return &worker_data->private_gray_queue;
422 SGEN_ASSERT (0, default_gray_queue, "Why don't we have a default gray queue when we're not running in a worker thread?");
423 return default_gray_queue;
427 gray_queue_redirect (SgenGrayQueue *queue)
429 SGEN_ASSERT (0, concurrent_collection_in_progress, "Where are we redirecting the gray queue to, without a concurrent collection?");
431 sgen_workers_take_from_queue (queue);
435 sgen_scan_area_with_callback (char *start, char *end, IterateObjectCallbackFunc callback, void *data, gboolean allow_flags, gboolean fail_on_canaries)
437 while (start < end) {
441 if (!*(void**)start) {
442 start += sizeof (void*); /* should be ALLOC_ALIGN, really */
447 if (!(obj = (char *)SGEN_OBJECT_IS_FORWARDED (start)))
453 if (!sgen_client_object_is_array_fill ((GCObject*)obj)) {
454 CHECK_CANARY_FOR_OBJECT ((GCObject*)obj, fail_on_canaries);
455 size = ALIGN_UP (safe_object_get_size ((GCObject*)obj));
456 callback ((GCObject*)obj, size, data);
457 CANARIFY_SIZE (size);
459 size = ALIGN_UP (safe_object_get_size ((GCObject*)obj));
467 * sgen_add_to_global_remset:
469 * The global remset contains locations which point into newspace after
470 * a minor collection. This can happen if the objects they point to are pinned.
472 * LOCKING: If called from a parallel collector, the global remset
473 * lock must be held. For serial collectors that is not necessary.
476 sgen_add_to_global_remset (gpointer ptr, GCObject *obj)
478 SGEN_ASSERT (5, sgen_ptr_in_nursery (obj), "Target pointer of global remset must be in the nursery");
480 HEAVY_STAT (++stat_wbarrier_add_to_global_remset);
482 if (!major_collector.is_concurrent) {
483 SGEN_ASSERT (5, current_collection_generation != -1, "Global remsets can only be added during collections");
485 if (current_collection_generation == -1)
486 SGEN_ASSERT (5, sgen_concurrent_collection_in_progress (), "Global remsets outside of collection pauses can only be added by the concurrent collector");
489 if (!object_is_pinned (obj))
490 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");
491 else if (sgen_cement_lookup_or_register (obj))
494 remset.record_pointer (ptr);
496 sgen_pin_stats_register_global_remset (obj);
498 SGEN_LOG (8, "Adding global remset for %p", ptr);
499 binary_protocol_global_remset (ptr, obj, (gpointer)SGEN_LOAD_VTABLE (obj));
503 * sgen_drain_gray_stack:
505 * Scan objects in the gray stack until the stack is empty. This should be called
506 * frequently after each object is copied, to achieve better locality and cache
511 sgen_drain_gray_stack (ScanCopyContext ctx)
513 SGEN_ASSERT (0, ctx.ops->drain_gray_stack, "Why do we have a scan/copy context with a missing drain gray stack function?");
515 return ctx.ops->drain_gray_stack (ctx.queue);
519 * Addresses in the pin queue are already sorted. This function finds
520 * the object header for each address and pins the object. The
521 * addresses must be inside the nursery section. The (start of the)
522 * address array is overwritten with the addresses of the actually
523 * pinned objects. Return the number of pinned objects.
526 pin_objects_from_nursery_pin_queue (gboolean do_scan_objects, ScanCopyContext ctx)
528 GCMemSection *section = nursery_section;
529 void **start = sgen_pinning_get_entry (section->pin_queue_first_entry);
530 void **end = sgen_pinning_get_entry (section->pin_queue_last_entry);
531 void *start_nursery = section->data;
532 void *end_nursery = section->next_data;
537 void *pinning_front = start_nursery;
539 void **definitely_pinned = start;
540 ScanObjectFunc scan_func = ctx.ops->scan_object;
541 SgenGrayQueue *queue = ctx.queue;
543 sgen_nursery_allocator_prepare_for_pinning ();
545 while (start < end) {
546 GCObject *obj_to_pin = NULL;
547 size_t obj_to_pin_size = 0;
552 SGEN_ASSERT (0, addr >= start_nursery && addr < end_nursery, "Potential pinning address out of range");
553 SGEN_ASSERT (0, addr >= last, "Pin queue not sorted");
560 SGEN_LOG (5, "Considering pinning addr %p", addr);
561 /* We've already processed everything up to pinning_front. */
562 if (addr < pinning_front) {
568 * Find the closest scan start <= addr. We might search backward in the
569 * scan_starts array because entries might be NULL. In the worst case we
570 * start at start_nursery.
572 idx = ((char*)addr - (char*)section->data) / SCAN_START_SIZE;
573 SGEN_ASSERT (0, idx < section->num_scan_start, "Scan start index out of range");
574 search_start = (void*)section->scan_starts [idx];
575 if (!search_start || search_start > addr) {
578 search_start = section->scan_starts [idx];
579 if (search_start && search_start <= addr)
582 if (!search_start || search_start > addr)
583 search_start = start_nursery;
587 * If the pinning front is closer than the scan start we found, start
588 * searching at the front.
590 if (search_start < pinning_front)
591 search_start = pinning_front;
594 * Now addr should be in an object a short distance from search_start.
596 * search_start must point to zeroed mem or point to an object.
599 size_t obj_size, canarified_obj_size;
602 if (!*(void**)search_start) {
603 search_start = (void*)ALIGN_UP ((mword)search_start + sizeof (gpointer));
604 /* The loop condition makes sure we don't overrun addr. */
608 canarified_obj_size = obj_size = ALIGN_UP (safe_object_get_size ((GCObject*)search_start));
611 * Filler arrays are marked by an invalid sync word. We don't
612 * consider them for pinning. They are not delimited by canaries,
615 if (!sgen_client_object_is_array_fill ((GCObject*)search_start)) {
616 CHECK_CANARY_FOR_OBJECT (search_start, TRUE);
617 CANARIFY_SIZE (canarified_obj_size);
619 if (addr >= search_start && (char*)addr < (char*)search_start + obj_size) {
620 /* This is the object we're looking for. */
621 obj_to_pin = (GCObject*)search_start;
622 obj_to_pin_size = canarified_obj_size;
627 /* Skip to the next object */
628 search_start = (void*)((char*)search_start + canarified_obj_size);
629 } while (search_start <= addr);
631 /* We've searched past the address we were looking for. */
633 pinning_front = search_start;
634 goto next_pin_queue_entry;
638 * We've found an object to pin. It might still be a dummy array, but we
639 * can advance the pinning front in any case.
641 pinning_front = (char*)obj_to_pin + obj_to_pin_size;
644 * If this is a dummy array marking the beginning of a nursery
645 * fragment, we don't pin it.
647 if (sgen_client_object_is_array_fill (obj_to_pin))
648 goto next_pin_queue_entry;
651 * Finally - pin the object!
653 desc = sgen_obj_get_descriptor_safe (obj_to_pin);
654 if (do_scan_objects) {
655 scan_func (obj_to_pin, desc, queue);
657 SGEN_LOG (4, "Pinned object %p, vtable %p (%s), count %d\n",
658 obj_to_pin, *(void**)obj_to_pin, sgen_client_vtable_get_name (SGEN_LOAD_VTABLE (obj_to_pin)), count);
659 binary_protocol_pin (obj_to_pin,
660 (gpointer)LOAD_VTABLE (obj_to_pin),
661 safe_object_get_size (obj_to_pin));
663 pin_object (obj_to_pin);
664 GRAY_OBJECT_ENQUEUE_SERIAL (queue, obj_to_pin, desc);
665 sgen_pin_stats_register_object (obj_to_pin, GENERATION_NURSERY);
666 definitely_pinned [count] = obj_to_pin;
669 if (concurrent_collection_in_progress)
670 sgen_pinning_register_pinned_in_nursery (obj_to_pin);
672 next_pin_queue_entry:
676 sgen_client_nursery_objects_pinned (definitely_pinned, count);
677 stat_pinned_objects += count;
682 pin_objects_in_nursery (gboolean do_scan_objects, ScanCopyContext ctx)
686 if (nursery_section->pin_queue_first_entry == nursery_section->pin_queue_last_entry)
689 reduced_to = pin_objects_from_nursery_pin_queue (do_scan_objects, ctx);
690 nursery_section->pin_queue_last_entry = nursery_section->pin_queue_first_entry + reduced_to;
694 * This function is only ever called (via `collector_pin_object()` in `sgen-copy-object.h`)
695 * when we can't promote an object because we're out of memory.
698 sgen_pin_object (GCObject *object, SgenGrayQueue *queue)
700 SGEN_ASSERT (0, sgen_ptr_in_nursery (object), "We're only supposed to use this for pinning nursery objects when out of memory.");
703 * All pinned objects are assumed to have been staged, so we need to stage as well.
704 * Also, the count of staged objects shows that "late pinning" happened.
706 sgen_pin_stage_ptr (object);
708 SGEN_PIN_OBJECT (object);
709 binary_protocol_pin (object, (gpointer)LOAD_VTABLE (object), safe_object_get_size (object));
712 sgen_pin_stats_register_object (object, GENERATION_NURSERY);
714 GRAY_OBJECT_ENQUEUE_SERIAL (queue, object, sgen_obj_get_descriptor_safe (object));
717 /* Sort the addresses in array in increasing order.
718 * Done using a by-the book heap sort. Which has decent and stable performance, is pretty cache efficient.
721 sgen_sort_addresses (void **array, size_t size)
726 for (i = 1; i < size; ++i) {
729 size_t parent = (child - 1) / 2;
731 if (array [parent] >= array [child])
734 tmp = array [parent];
735 array [parent] = array [child];
742 for (i = size - 1; i > 0; --i) {
745 array [i] = array [0];
751 while (root * 2 + 1 <= end) {
752 size_t child = root * 2 + 1;
754 if (child < end && array [child] < array [child + 1])
756 if (array [root] >= array [child])
760 array [root] = array [child];
769 * Scan the memory between start and end and queue values which could be pointers
770 * to the area between start_nursery and end_nursery for later consideration.
771 * Typically used for thread stacks.
774 sgen_conservatively_pin_objects_from (void **start, void **end, void *start_nursery, void *end_nursery, int pin_type)
778 SGEN_ASSERT (0, ((mword)start & (SIZEOF_VOID_P - 1)) == 0, "Why are we scanning for references in unaligned memory ?");
780 #if defined(VALGRIND_MAKE_MEM_DEFINED_IF_ADDRESSABLE) && !defined(_WIN64)
781 VALGRIND_MAKE_MEM_DEFINED_IF_ADDRESSABLE (start, (char*)end - (char*)start);
784 while (start < end) {
786 * *start can point to the middle of an object
787 * note: should we handle pointing at the end of an object?
788 * pinning in C# code disallows pointing at the end of an object
789 * but there is some small chance that an optimizing C compiler
790 * may keep the only reference to an object by pointing
791 * at the end of it. We ignore this small chance for now.
792 * Pointers to the end of an object are indistinguishable
793 * from pointers to the start of the next object in memory
794 * so if we allow that we'd need to pin two objects...
795 * We queue the pointer in an array, the
796 * array will then be sorted and uniqued. This way
797 * we can coalesce several pinning pointers and it should
798 * be faster since we'd do a memory scan with increasing
799 * addresses. Note: we can align the address to the allocation
800 * alignment, so the unique process is more effective.
802 mword addr = (mword)*start;
803 addr &= ~(ALLOC_ALIGN - 1);
804 if (addr >= (mword)start_nursery && addr < (mword)end_nursery) {
805 SGEN_LOG (6, "Pinning address %p from %p", (void*)addr, start);
806 sgen_pin_stage_ptr ((void*)addr);
807 binary_protocol_pin_stage (start, (void*)addr);
808 sgen_pin_stats_register_address ((char*)addr, pin_type);
814 SGEN_LOG (7, "found %d potential pinned heap pointers", count);
818 * The first thing we do in a collection is to identify pinned objects.
819 * This function considers all the areas of memory that need to be
820 * conservatively scanned.
823 pin_from_roots (void *start_nursery, void *end_nursery, ScanCopyContext ctx)
827 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);
828 /* objects pinned from the API are inside these roots */
829 SGEN_HASH_TABLE_FOREACH (&roots_hash [ROOT_TYPE_PINNED], void **, start_root, RootRecord *, root) {
830 SGEN_LOG (6, "Pinned roots %p-%p", start_root, root->end_root);
831 sgen_conservatively_pin_objects_from (start_root, (void**)root->end_root, start_nursery, end_nursery, PIN_TYPE_OTHER);
832 } SGEN_HASH_TABLE_FOREACH_END;
833 /* now deal with the thread stacks
834 * in the future we should be able to conservatively scan only:
835 * *) the cpu registers
836 * *) the unmanaged stack frames
837 * *) the _last_ managed stack frame
838 * *) pointers slots in managed frames
840 sgen_client_scan_thread_data (start_nursery, end_nursery, FALSE, ctx);
844 single_arg_user_copy_or_mark (GCObject **obj, void *gc_data)
846 ScanCopyContext *ctx = (ScanCopyContext *)gc_data;
847 ctx->ops->copy_or_mark_object (obj, ctx->queue);
851 * The memory area from start_root to end_root contains pointers to objects.
852 * Their position is precisely described by @desc (this means that the pointer
853 * can be either NULL or the pointer to the start of an object).
854 * This functions copies them to to_space updates them.
856 * This function is not thread-safe!
859 precisely_scan_objects_from (void** start_root, void** end_root, char* n_start, char *n_end, SgenDescriptor desc, ScanCopyContext ctx)
861 CopyOrMarkObjectFunc copy_func = ctx.ops->copy_or_mark_object;
862 ScanPtrFieldFunc scan_field_func = ctx.ops->scan_ptr_field;
863 SgenGrayQueue *queue = ctx.queue;
865 switch (desc & ROOT_DESC_TYPE_MASK) {
866 case ROOT_DESC_BITMAP:
867 desc >>= ROOT_DESC_TYPE_SHIFT;
869 if ((desc & 1) && *start_root) {
870 copy_func ((GCObject**)start_root, queue);
871 SGEN_LOG (9, "Overwrote root at %p with %p", start_root, *start_root);
877 case ROOT_DESC_COMPLEX: {
878 gsize *bitmap_data = (gsize *)sgen_get_complex_descriptor_bitmap (desc);
879 gsize bwords = (*bitmap_data) - 1;
880 void **start_run = start_root;
882 while (bwords-- > 0) {
883 gsize bmap = *bitmap_data++;
884 void **objptr = start_run;
886 if ((bmap & 1) && *objptr) {
887 copy_func ((GCObject**)objptr, queue);
888 SGEN_LOG (9, "Overwrote root at %p with %p", objptr, *objptr);
893 start_run += GC_BITS_PER_WORD;
897 case ROOT_DESC_VECTOR: {
900 for (p = start_root; p < end_root; p++) {
902 scan_field_func (NULL, (GCObject**)p, queue);
906 case ROOT_DESC_USER: {
907 SgenUserRootMarkFunc marker = sgen_get_user_descriptor_func (desc);
908 marker (start_root, single_arg_user_copy_or_mark, &ctx);
911 case ROOT_DESC_RUN_LEN:
912 g_assert_not_reached ();
914 g_assert_not_reached ();
919 reset_heap_boundaries (void)
921 lowest_heap_address = ~(mword)0;
922 highest_heap_address = 0;
926 sgen_update_heap_boundaries (mword low, mword high)
931 old = lowest_heap_address;
934 } while (SGEN_CAS_PTR ((gpointer*)&lowest_heap_address, (gpointer)low, (gpointer)old) != (gpointer)old);
937 old = highest_heap_address;
940 } while (SGEN_CAS_PTR ((gpointer*)&highest_heap_address, (gpointer)high, (gpointer)old) != (gpointer)old);
944 * Allocate and setup the data structures needed to be able to allocate objects
945 * in the nursery. The nursery is stored in nursery_section.
950 GCMemSection *section;
957 SGEN_LOG (2, "Allocating nursery size: %zu", (size_t)sgen_nursery_size);
958 /* later we will alloc a larger area for the nursery but only activate
959 * what we need. The rest will be used as expansion if we have too many pinned
960 * objects in the existing nursery.
962 /* FIXME: handle OOM */
963 section = (GCMemSection *)sgen_alloc_internal (INTERNAL_MEM_SECTION);
965 alloc_size = sgen_nursery_size;
967 /* If there isn't enough space even for the nursery we should simply abort. */
968 g_assert (sgen_memgov_try_alloc_space (alloc_size, SPACE_NURSERY));
970 data = (char *)major_collector.alloc_heap (alloc_size, alloc_size, DEFAULT_NURSERY_BITS);
971 sgen_update_heap_boundaries ((mword)data, (mword)(data + sgen_nursery_size));
972 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 ());
973 section->data = section->next_data = data;
974 section->size = alloc_size;
975 section->end_data = data + sgen_nursery_size;
976 scan_starts = (alloc_size + SCAN_START_SIZE - 1) / SCAN_START_SIZE;
977 section->scan_starts = (char **)sgen_alloc_internal_dynamic (sizeof (char*) * scan_starts, INTERNAL_MEM_SCAN_STARTS, TRUE);
978 section->num_scan_start = scan_starts;
980 nursery_section = section;
982 sgen_nursery_allocator_set_nursery_bounds (data, data + sgen_nursery_size);
986 mono_gc_get_logfile (void)
988 return gc_debug_file;
992 mono_gc_params_set (const char* options)
994 if (gc_params_options)
995 g_free (gc_params_options);
997 gc_params_options = g_strdup (options);
1001 mono_gc_debug_set (const char* options)
1003 if (gc_debug_options)
1004 g_free (gc_debug_options);
1006 gc_debug_options = g_strdup (options);
1010 scan_finalizer_entries (SgenPointerQueue *fin_queue, ScanCopyContext ctx)
1012 CopyOrMarkObjectFunc copy_func = ctx.ops->copy_or_mark_object;
1013 SgenGrayQueue *queue = ctx.queue;
1016 for (i = 0; i < fin_queue->next_slot; ++i) {
1017 GCObject *obj = (GCObject *)fin_queue->data [i];
1020 SGEN_LOG (5, "Scan of fin ready object: %p (%s)\n", obj, sgen_client_vtable_get_name (SGEN_LOAD_VTABLE (obj)));
1021 copy_func ((GCObject**)&fin_queue->data [i], queue);
1026 generation_name (int generation)
1028 switch (generation) {
1029 case GENERATION_NURSERY: return "nursery";
1030 case GENERATION_OLD: return "old";
1031 default: g_assert_not_reached ();
1036 sgen_generation_name (int generation)
1038 return generation_name (generation);
1042 finish_gray_stack (int generation, ScanCopyContext ctx)
1046 int done_with_ephemerons, ephemeron_rounds = 0;
1047 char *start_addr = generation == GENERATION_NURSERY ? sgen_get_nursery_start () : NULL;
1048 char *end_addr = generation == GENERATION_NURSERY ? sgen_get_nursery_end () : (char*)-1;
1049 SgenGrayQueue *queue = ctx.queue;
1051 binary_protocol_finish_gray_stack_start (sgen_timestamp (), generation);
1053 * We copied all the reachable objects. Now it's the time to copy
1054 * the objects that were not referenced by the roots, but by the copied objects.
1055 * we built a stack of objects pointed to by gray_start: they are
1056 * additional roots and we may add more items as we go.
1057 * We loop until gray_start == gray_objects which means no more objects have
1058 * been added. Note this is iterative: no recursion is involved.
1059 * We need to walk the LO list as well in search of marked big objects
1060 * (use a flag since this is needed only on major collections). We need to loop
1061 * here as well, so keep a counter of marked LO (increasing it in copy_object).
1062 * To achieve better cache locality and cache usage, we drain the gray stack
1063 * frequently, after each object is copied, and just finish the work here.
1065 sgen_drain_gray_stack (ctx);
1067 SGEN_LOG (2, "%s generation done", generation_name (generation));
1070 Reset bridge data, we might have lingering data from a previous collection if this is a major
1071 collection trigged by minor overflow.
1073 We must reset the gathered bridges since their original block might be evacuated due to major
1074 fragmentation in the meanwhile and the bridge code should not have to deal with that.
1076 if (sgen_client_bridge_need_processing ())
1077 sgen_client_bridge_reset_data ();
1080 * Mark all strong toggleref objects. This must be done before we walk ephemerons or finalizers
1081 * to ensure they see the full set of live objects.
1083 sgen_client_mark_togglerefs (start_addr, end_addr, ctx);
1086 * Walk the ephemeron tables marking all values with reachable keys. This must be completely done
1087 * before processing finalizable objects and non-tracking weak links to avoid finalizing/clearing
1088 * objects that are in fact reachable.
1090 done_with_ephemerons = 0;
1092 done_with_ephemerons = sgen_client_mark_ephemerons (ctx);
1093 sgen_drain_gray_stack (ctx);
1095 } while (!done_with_ephemerons);
1097 if (sgen_client_bridge_need_processing ()) {
1098 /*Make sure the gray stack is empty before we process bridge objects so we get liveness right*/
1099 sgen_drain_gray_stack (ctx);
1100 sgen_collect_bridge_objects (generation, ctx);
1101 if (generation == GENERATION_OLD)
1102 sgen_collect_bridge_objects (GENERATION_NURSERY, ctx);
1105 Do the first bridge step here, as the collector liveness state will become useless after that.
1107 An important optimization is to only proccess the possibly dead part of the object graph and skip
1108 over all live objects as we transitively know everything they point must be alive too.
1110 The above invariant is completely wrong if we let the gray queue be drained and mark/copy everything.
1112 This has the unfortunate side effect of making overflow collections perform the first step twice, but
1113 given we now have heuristics that perform major GC in anticipation of minor overflows this should not
1116 sgen_client_bridge_processing_stw_step ();
1120 Make sure we drain the gray stack before processing disappearing links and finalizers.
1121 If we don't make sure it is empty we might wrongly see a live object as dead.
1123 sgen_drain_gray_stack (ctx);
1126 We must clear weak links that don't track resurrection before processing object ready for
1127 finalization so they can be cleared before that.
1129 sgen_null_link_in_range (generation, ctx, FALSE);
1130 if (generation == GENERATION_OLD)
1131 sgen_null_link_in_range (GENERATION_NURSERY, ctx, FALSE);
1134 /* walk the finalization queue and move also the objects that need to be
1135 * finalized: use the finalized objects as new roots so the objects they depend
1136 * on are also not reclaimed. As with the roots above, only objects in the nursery
1137 * are marked/copied.
1139 sgen_finalize_in_range (generation, ctx);
1140 if (generation == GENERATION_OLD)
1141 sgen_finalize_in_range (GENERATION_NURSERY, ctx);
1142 /* drain the new stack that might have been created */
1143 SGEN_LOG (6, "Precise scan of gray area post fin");
1144 sgen_drain_gray_stack (ctx);
1147 * This must be done again after processing finalizable objects since CWL slots are cleared only after the key is finalized.
1149 done_with_ephemerons = 0;
1151 done_with_ephemerons = sgen_client_mark_ephemerons (ctx);
1152 sgen_drain_gray_stack (ctx);
1154 } while (!done_with_ephemerons);
1156 sgen_client_clear_unreachable_ephemerons (ctx);
1159 * We clear togglerefs only after all possible chances of revival are done.
1160 * This is semantically more inline with what users expect and it allows for
1161 * user finalizers to correctly interact with TR objects.
1163 sgen_client_clear_togglerefs (start_addr, end_addr, ctx);
1166 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);
1169 * handle disappearing links
1170 * Note we do this after checking the finalization queue because if an object
1171 * survives (at least long enough to be finalized) we don't clear the link.
1172 * This also deals with a possible issue with the monitor reclamation: with the Boehm
1173 * GC a finalized object my lose the monitor because it is cleared before the finalizer is
1176 g_assert (sgen_gray_object_queue_is_empty (queue));
1178 sgen_null_link_in_range (generation, ctx, TRUE);
1179 if (generation == GENERATION_OLD)
1180 sgen_null_link_in_range (GENERATION_NURSERY, ctx, TRUE);
1181 if (sgen_gray_object_queue_is_empty (queue))
1183 sgen_drain_gray_stack (ctx);
1186 g_assert (sgen_gray_object_queue_is_empty (queue));
1188 binary_protocol_finish_gray_stack_end (sgen_timestamp (), generation);
1192 sgen_check_section_scan_starts (GCMemSection *section)
1195 for (i = 0; i < section->num_scan_start; ++i) {
1196 if (section->scan_starts [i]) {
1197 mword size = safe_object_get_size ((GCObject*) section->scan_starts [i]);
1198 SGEN_ASSERT (0, size >= SGEN_CLIENT_MINIMUM_OBJECT_SIZE && size <= MAX_SMALL_OBJ_SIZE, "Weird object size at scan starts.");
1204 check_scan_starts (void)
1206 if (!do_scan_starts_check)
1208 sgen_check_section_scan_starts (nursery_section);
1209 major_collector.check_scan_starts ();
1213 scan_from_registered_roots (char *addr_start, char *addr_end, int root_type, ScanCopyContext ctx)
1217 SGEN_HASH_TABLE_FOREACH (&roots_hash [root_type], void **, start_root, RootRecord *, root) {
1218 SGEN_LOG (6, "Precise root scan %p-%p (desc: %p)", start_root, root->end_root, (void*)root->root_desc);
1219 precisely_scan_objects_from (start_root, (void**)root->end_root, addr_start, addr_end, root->root_desc, ctx);
1220 } SGEN_HASH_TABLE_FOREACH_END;
1226 static gboolean inited = FALSE;
1231 mono_counters_register ("Collection max time", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME | MONO_COUNTER_MONOTONIC, &time_max);
1233 mono_counters_register ("Minor fragment clear", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_pre_collection_fragment_clear);
1234 mono_counters_register ("Minor pinning", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_pinning);
1235 mono_counters_register ("Minor scan remembered set", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_scan_remsets);
1236 mono_counters_register ("Minor scan pinned", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_scan_pinned);
1237 mono_counters_register ("Minor scan roots", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_scan_roots);
1238 mono_counters_register ("Minor fragment creation", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_minor_fragment_creation);
1240 mono_counters_register ("Major fragment clear", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_pre_collection_fragment_clear);
1241 mono_counters_register ("Major pinning", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_pinning);
1242 mono_counters_register ("Major scan pinned", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_scan_pinned);
1243 mono_counters_register ("Major scan roots", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_scan_roots);
1244 mono_counters_register ("Major scan mod union", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_scan_mod_union);
1245 mono_counters_register ("Major finish gray stack", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_finish_gray_stack);
1246 mono_counters_register ("Major free big objects", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_free_bigobjs);
1247 mono_counters_register ("Major LOS sweep", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_los_sweep);
1248 mono_counters_register ("Major sweep", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_sweep);
1249 mono_counters_register ("Major fragment creation", MONO_COUNTER_GC | MONO_COUNTER_ULONG | MONO_COUNTER_TIME, &time_major_fragment_creation);
1251 mono_counters_register ("Number of pinned objects", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_pinned_objects);
1253 #ifdef HEAVY_STATISTICS
1254 mono_counters_register ("WBarrier remember pointer", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_wbarrier_add_to_global_remset);
1255 mono_counters_register ("WBarrier arrayref copy", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_wbarrier_arrayref_copy);
1256 mono_counters_register ("WBarrier generic store called", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_wbarrier_generic_store);
1257 mono_counters_register ("WBarrier generic atomic store called", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_wbarrier_generic_store_atomic);
1258 mono_counters_register ("WBarrier set root", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_wbarrier_set_root);
1260 mono_counters_register ("# objects allocated degraded", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_objects_alloced_degraded);
1261 mono_counters_register ("bytes allocated degraded", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_bytes_alloced_degraded);
1263 mono_counters_register ("# copy_object() called (nursery)", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_copy_object_called_nursery);
1264 mono_counters_register ("# objects copied (nursery)", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_objects_copied_nursery);
1265 mono_counters_register ("# copy_object() called (major)", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_copy_object_called_major);
1266 mono_counters_register ("# objects copied (major)", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_objects_copied_major);
1268 mono_counters_register ("# scan_object() called (nursery)", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_scan_object_called_nursery);
1269 mono_counters_register ("# scan_object() called (major)", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_scan_object_called_major);
1271 mono_counters_register ("Slots allocated in vain", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_slots_allocated_in_vain);
1273 mono_counters_register ("# nursery copy_object() failed from space", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_nursery_copy_object_failed_from_space);
1274 mono_counters_register ("# nursery copy_object() failed forwarded", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_nursery_copy_object_failed_forwarded);
1275 mono_counters_register ("# nursery copy_object() failed pinned", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_nursery_copy_object_failed_pinned);
1276 mono_counters_register ("# nursery copy_object() failed to space", MONO_COUNTER_GC | MONO_COUNTER_ULONG, &stat_nursery_copy_object_failed_to_space);
1278 sgen_nursery_allocator_init_heavy_stats ();
1286 reset_pinned_from_failed_allocation (void)
1288 bytes_pinned_from_failed_allocation = 0;
1292 sgen_set_pinned_from_failed_allocation (mword objsize)
1294 bytes_pinned_from_failed_allocation += objsize;
1298 sgen_collection_is_concurrent (void)
1300 switch (current_collection_generation) {
1301 case GENERATION_NURSERY:
1303 case GENERATION_OLD:
1304 return concurrent_collection_in_progress;
1306 g_error ("Invalid current generation %d", current_collection_generation);
1312 sgen_concurrent_collection_in_progress (void)
1314 return concurrent_collection_in_progress;
1318 SgenThreadPoolJob job;
1319 SgenObjectOperations *ops;
1320 SgenGrayQueue *gc_thread_gray_queue;
1328 static ScanCopyContext
1329 scan_copy_context_for_scan_job (void *worker_data_untyped, ScanJob *job)
1331 WorkerData *worker_data = (WorkerData *)worker_data_untyped;
1333 return CONTEXT_FROM_OBJECT_OPERATIONS (job->ops, sgen_workers_get_job_gray_queue (worker_data, job->gc_thread_gray_queue));
1337 job_remembered_set_scan (void *worker_data_untyped, SgenThreadPoolJob *job)
1339 remset.scan_remsets (scan_copy_context_for_scan_job (worker_data_untyped, (ScanJob*)job));
1347 } ScanFromRegisteredRootsJob;
1350 job_scan_from_registered_roots (void *worker_data_untyped, SgenThreadPoolJob *job)
1352 ScanFromRegisteredRootsJob *job_data = (ScanFromRegisteredRootsJob*)job;
1353 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, &job_data->scan_job);
1355 scan_from_registered_roots (job_data->heap_start, job_data->heap_end, job_data->root_type, ctx);
1362 } ScanThreadDataJob;
1365 job_scan_thread_data (void *worker_data_untyped, SgenThreadPoolJob *job)
1367 ScanThreadDataJob *job_data = (ScanThreadDataJob*)job;
1368 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, &job_data->scan_job);
1370 sgen_client_scan_thread_data (job_data->heap_start, job_data->heap_end, TRUE, ctx);
1375 SgenPointerQueue *queue;
1376 } ScanFinalizerEntriesJob;
1379 job_scan_finalizer_entries (void *worker_data_untyped, SgenThreadPoolJob *job)
1381 ScanFinalizerEntriesJob *job_data = (ScanFinalizerEntriesJob*)job;
1382 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, &job_data->scan_job);
1384 scan_finalizer_entries (job_data->queue, ctx);
1388 job_scan_major_mod_union_card_table (void *worker_data_untyped, SgenThreadPoolJob *job)
1390 ParallelScanJob *job_data = (ParallelScanJob*)job;
1391 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, (ScanJob*)job_data);
1393 g_assert (concurrent_collection_in_progress);
1394 major_collector.scan_card_table (CARDTABLE_SCAN_MOD_UNION, ctx, job_data->job_index, sgen_workers_get_job_split_count ());
1398 job_scan_los_mod_union_card_table (void *worker_data_untyped, SgenThreadPoolJob *job)
1400 ParallelScanJob *job_data = (ParallelScanJob*)job;
1401 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, (ScanJob*)job_data);
1403 g_assert (concurrent_collection_in_progress);
1404 sgen_los_scan_card_table (CARDTABLE_SCAN_MOD_UNION, ctx, job_data->job_index, sgen_workers_get_job_split_count ());
1408 job_major_mod_union_preclean (void *worker_data_untyped, SgenThreadPoolJob *job)
1410 ParallelScanJob *job_data = (ParallelScanJob*)job;
1411 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, (ScanJob*)job_data);
1413 g_assert (concurrent_collection_in_progress);
1415 major_collector.scan_card_table (CARDTABLE_SCAN_MOD_UNION_PRECLEAN, ctx, job_data->job_index, sgen_workers_get_job_split_count ());
1419 job_los_mod_union_preclean (void *worker_data_untyped, SgenThreadPoolJob *job)
1421 ParallelScanJob *job_data = (ParallelScanJob*)job;
1422 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, (ScanJob*)job_data);
1424 g_assert (concurrent_collection_in_progress);
1426 sgen_los_scan_card_table (CARDTABLE_SCAN_MOD_UNION_PRECLEAN, ctx, job_data->job_index, sgen_workers_get_job_split_count ());
1430 job_scan_last_pinned (void *worker_data_untyped, SgenThreadPoolJob *job)
1432 ScanJob *job_data = (ScanJob*)job;
1433 ScanCopyContext ctx = scan_copy_context_for_scan_job (worker_data_untyped, job_data);
1435 g_assert (concurrent_collection_in_progress);
1437 sgen_scan_pin_queue_objects (ctx);
1441 workers_finish_callback (void)
1443 ParallelScanJob *psj;
1445 int split_count = sgen_workers_get_job_split_count ();
1447 /* Mod union preclean jobs */
1448 for (i = 0; i < split_count; i++) {
1449 psj = (ParallelScanJob*)sgen_thread_pool_job_alloc ("preclean major mod union cardtable", job_major_mod_union_preclean, sizeof (ParallelScanJob));
1450 psj->scan_job.ops = sgen_workers_get_idle_func_object_ops ();
1451 psj->scan_job.gc_thread_gray_queue = NULL;
1453 sgen_workers_enqueue_job (&psj->scan_job.job, TRUE);
1456 for (i = 0; i < split_count; i++) {
1457 psj = (ParallelScanJob*)sgen_thread_pool_job_alloc ("preclean los mod union cardtable", job_los_mod_union_preclean, sizeof (ParallelScanJob));
1458 psj->scan_job.ops = sgen_workers_get_idle_func_object_ops ();
1459 psj->scan_job.gc_thread_gray_queue = NULL;
1461 sgen_workers_enqueue_job (&psj->scan_job.job, TRUE);
1464 sj = (ScanJob*)sgen_thread_pool_job_alloc ("scan last pinned", job_scan_last_pinned, sizeof (ScanJob));
1465 sj->ops = sgen_workers_get_idle_func_object_ops ();
1466 sj->gc_thread_gray_queue = NULL;
1467 sgen_workers_enqueue_job (&sj->job, TRUE);
1471 init_gray_queue (SgenGrayQueue *gc_thread_gray_queue, gboolean use_workers)
1474 sgen_workers_init_distribute_gray_queue ();
1475 sgen_gray_object_queue_init (gc_thread_gray_queue, NULL, TRUE);
1479 enqueue_scan_from_roots_jobs (SgenGrayQueue *gc_thread_gray_queue, char *heap_start, char *heap_end, SgenObjectOperations *ops, gboolean enqueue)
1481 ScanFromRegisteredRootsJob *scrrj;
1482 ScanThreadDataJob *stdj;
1483 ScanFinalizerEntriesJob *sfej;
1485 /* registered roots, this includes static fields */
1487 scrrj = (ScanFromRegisteredRootsJob*)sgen_thread_pool_job_alloc ("scan from registered roots normal", job_scan_from_registered_roots, sizeof (ScanFromRegisteredRootsJob));
1488 scrrj->scan_job.ops = ops;
1489 scrrj->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1490 scrrj->heap_start = heap_start;
1491 scrrj->heap_end = heap_end;
1492 scrrj->root_type = ROOT_TYPE_NORMAL;
1493 sgen_workers_enqueue_job (&scrrj->scan_job.job, enqueue);
1495 if (current_collection_generation == GENERATION_OLD) {
1496 /* During minors we scan the cardtable for these roots instead */
1497 scrrj = (ScanFromRegisteredRootsJob*)sgen_thread_pool_job_alloc ("scan from registered roots wbarrier", job_scan_from_registered_roots, sizeof (ScanFromRegisteredRootsJob));
1498 scrrj->scan_job.ops = ops;
1499 scrrj->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1500 scrrj->heap_start = heap_start;
1501 scrrj->heap_end = heap_end;
1502 scrrj->root_type = ROOT_TYPE_WBARRIER;
1503 sgen_workers_enqueue_job (&scrrj->scan_job.job, enqueue);
1508 stdj = (ScanThreadDataJob*)sgen_thread_pool_job_alloc ("scan thread data", job_scan_thread_data, sizeof (ScanThreadDataJob));
1509 stdj->scan_job.ops = ops;
1510 stdj->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1511 stdj->heap_start = heap_start;
1512 stdj->heap_end = heap_end;
1513 sgen_workers_enqueue_job (&stdj->scan_job.job, enqueue);
1515 /* Scan the list of objects ready for finalization. */
1517 sfej = (ScanFinalizerEntriesJob*)sgen_thread_pool_job_alloc ("scan finalizer entries", job_scan_finalizer_entries, sizeof (ScanFinalizerEntriesJob));
1518 sfej->scan_job.ops = ops;
1519 sfej->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1520 sfej->queue = &fin_ready_queue;
1521 sgen_workers_enqueue_job (&sfej->scan_job.job, enqueue);
1523 sfej = (ScanFinalizerEntriesJob*)sgen_thread_pool_job_alloc ("scan critical finalizer entries", job_scan_finalizer_entries, sizeof (ScanFinalizerEntriesJob));
1524 sfej->scan_job.ops = ops;
1525 sfej->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1526 sfej->queue = &critical_fin_queue;
1527 sgen_workers_enqueue_job (&sfej->scan_job.job, enqueue);
1531 * Perform a nursery collection.
1533 * Return whether any objects were late-pinned due to being out of memory.
1536 collect_nursery (const char *reason, gboolean is_overflow, SgenGrayQueue *unpin_queue)
1538 gboolean needs_major;
1539 size_t max_garbage_amount;
1541 mword fragment_total;
1543 SgenGrayQueue gc_thread_gray_queue;
1544 SgenObjectOperations *object_ops;
1545 ScanCopyContext ctx;
1548 SGEN_TV_DECLARE (last_minor_collection_start_tv);
1549 SGEN_TV_DECLARE (last_minor_collection_end_tv);
1551 if (disable_minor_collections)
1554 TV_GETTIME (last_minor_collection_start_tv);
1555 atv = last_minor_collection_start_tv;
1557 binary_protocol_collection_begin (gc_stats.minor_gc_count, GENERATION_NURSERY);
1559 if (sgen_concurrent_collection_in_progress ())
1560 object_ops = &sgen_minor_collector.serial_ops_with_concurrent_major;
1562 object_ops = &sgen_minor_collector.serial_ops;
1564 if (do_verify_nursery || do_dump_nursery_content)
1565 sgen_debug_verify_nursery (do_dump_nursery_content);
1567 current_collection_generation = GENERATION_NURSERY;
1569 SGEN_ASSERT (0, !sgen_collection_is_concurrent (), "Why is the nursery collection concurrent?");
1571 reset_pinned_from_failed_allocation ();
1573 check_scan_starts ();
1575 sgen_nursery_alloc_prepare_for_minor ();
1579 nursery_next = sgen_nursery_alloc_get_upper_alloc_bound ();
1580 /* FIXME: optimize later to use the higher address where an object can be present */
1581 nursery_next = MAX (nursery_next, sgen_get_nursery_end ());
1583 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 ()));
1584 max_garbage_amount = nursery_next - sgen_get_nursery_start ();
1585 g_assert (nursery_section->size >= max_garbage_amount);
1587 /* world must be stopped already */
1589 time_minor_pre_collection_fragment_clear += TV_ELAPSED (atv, btv);
1591 sgen_client_pre_collection_checks ();
1593 nursery_section->next_data = nursery_next;
1595 major_collector.start_nursery_collection ();
1597 sgen_memgov_minor_collection_start ();
1599 init_gray_queue (&gc_thread_gray_queue, FALSE);
1600 ctx = CONTEXT_FROM_OBJECT_OPERATIONS (object_ops, &gc_thread_gray_queue);
1602 gc_stats.minor_gc_count ++;
1604 sgen_process_fin_stage_entries ();
1606 /* pin from pinned handles */
1607 sgen_init_pinning ();
1608 sgen_client_binary_protocol_mark_start (GENERATION_NURSERY);
1609 pin_from_roots (sgen_get_nursery_start (), nursery_next, ctx);
1610 /* pin cemented objects */
1611 sgen_pin_cemented_objects ();
1612 /* identify pinned objects */
1613 sgen_optimize_pin_queue ();
1614 sgen_pinning_setup_section (nursery_section);
1616 pin_objects_in_nursery (FALSE, ctx);
1617 sgen_pinning_trim_queue_to_section (nursery_section);
1619 if (remset_consistency_checks)
1620 sgen_check_remset_consistency ();
1622 if (whole_heap_check_before_collection) {
1623 sgen_clear_nursery_fragments ();
1624 sgen_check_whole_heap (FALSE);
1628 time_minor_pinning += TV_ELAPSED (btv, atv);
1629 SGEN_LOG (2, "Finding pinned pointers: %zd in %lld usecs", sgen_get_pinned_count (), (long long)TV_ELAPSED (btv, atv));
1630 SGEN_LOG (4, "Start scan with %zd pinned objects", sgen_get_pinned_count ());
1632 sj = (ScanJob*)sgen_thread_pool_job_alloc ("scan remset", job_remembered_set_scan, sizeof (ScanJob));
1633 sj->ops = object_ops;
1634 sj->gc_thread_gray_queue = &gc_thread_gray_queue;
1635 sgen_workers_enqueue_job (&sj->job, FALSE);
1637 /* we don't have complete write barrier yet, so we scan all the old generation sections */
1639 time_minor_scan_remsets += TV_ELAPSED (atv, btv);
1640 SGEN_LOG (2, "Old generation scan: %lld usecs", (long long)TV_ELAPSED (atv, btv));
1642 sgen_pin_stats_report ();
1644 /* FIXME: Why do we do this at this specific, seemingly random, point? */
1645 sgen_client_collecting_minor (&fin_ready_queue, &critical_fin_queue);
1648 time_minor_scan_pinned += TV_ELAPSED (btv, atv);
1650 enqueue_scan_from_roots_jobs (&gc_thread_gray_queue, sgen_get_nursery_start (), nursery_next, object_ops, FALSE);
1653 time_minor_scan_roots += TV_ELAPSED (atv, btv);
1655 finish_gray_stack (GENERATION_NURSERY, ctx);
1658 time_minor_finish_gray_stack += TV_ELAPSED (btv, atv);
1659 sgen_client_binary_protocol_mark_end (GENERATION_NURSERY);
1661 if (objects_pinned) {
1662 sgen_optimize_pin_queue ();
1663 sgen_pinning_setup_section (nursery_section);
1667 * This is the latest point at which we can do this check, because
1668 * sgen_build_nursery_fragments() unpins nursery objects again.
1670 if (remset_consistency_checks)
1671 sgen_check_remset_consistency ();
1673 /* walk the pin_queue, build up the fragment list of free memory, unmark
1674 * pinned objects as we go, memzero() the empty fragments so they are ready for the
1677 sgen_client_binary_protocol_reclaim_start (GENERATION_NURSERY);
1678 fragment_total = sgen_build_nursery_fragments (nursery_section, unpin_queue);
1679 if (!fragment_total)
1682 /* Clear TLABs for all threads */
1683 sgen_clear_tlabs ();
1685 sgen_client_binary_protocol_reclaim_end (GENERATION_NURSERY);
1687 time_minor_fragment_creation += TV_ELAPSED (atv, btv);
1688 SGEN_LOG (2, "Fragment creation: %lld usecs, %lu bytes available", (long long)TV_ELAPSED (atv, btv), (unsigned long)fragment_total);
1690 if (remset_consistency_checks)
1691 sgen_check_major_refs ();
1693 major_collector.finish_nursery_collection ();
1695 TV_GETTIME (last_minor_collection_end_tv);
1696 gc_stats.minor_gc_time += TV_ELAPSED (last_minor_collection_start_tv, last_minor_collection_end_tv);
1698 sgen_debug_dump_heap ("minor", gc_stats.minor_gc_count - 1, NULL);
1700 /* prepare the pin queue for the next collection */
1701 sgen_finish_pinning ();
1702 if (sgen_have_pending_finalizers ()) {
1703 SGEN_LOG (4, "Finalizer-thread wakeup");
1704 sgen_client_finalize_notify ();
1706 sgen_pin_stats_reset ();
1707 /* clear cemented hash */
1708 sgen_cement_clear_below_threshold ();
1710 sgen_gray_object_queue_dispose (&gc_thread_gray_queue);
1712 check_scan_starts ();
1714 binary_protocol_flush_buffers (FALSE);
1716 sgen_memgov_minor_collection_end (reason, is_overflow);
1718 /*objects are late pinned because of lack of memory, so a major is a good call*/
1719 needs_major = objects_pinned > 0;
1720 current_collection_generation = -1;
1723 binary_protocol_collection_end (gc_stats.minor_gc_count - 1, GENERATION_NURSERY, 0, 0);
1725 if (check_nursery_objects_pinned && !sgen_minor_collector.is_split)
1726 sgen_check_nursery_objects_pinned (unpin_queue != NULL);
1732 COPY_OR_MARK_FROM_ROOTS_SERIAL,
1733 COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT,
1734 COPY_OR_MARK_FROM_ROOTS_FINISH_CONCURRENT
1735 } CopyOrMarkFromRootsMode;
1738 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)
1743 /* FIXME: only use these values for the precise scan
1744 * note that to_space pointers should be excluded anyway...
1746 char *heap_start = NULL;
1747 char *heap_end = (char*)-1;
1748 ScanCopyContext ctx = CONTEXT_FROM_OBJECT_OPERATIONS (object_ops_nopar, gc_thread_gray_queue);
1749 gboolean concurrent = mode != COPY_OR_MARK_FROM_ROOTS_SERIAL;
1751 SGEN_ASSERT (0, !!concurrent == !!concurrent_collection_in_progress, "We've been called with the wrong mode.");
1753 if (mode == COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT) {
1754 /*This cleans up unused fragments */
1755 sgen_nursery_allocator_prepare_for_pinning ();
1757 if (do_concurrent_checks)
1758 sgen_debug_check_nursery_is_clean ();
1760 /* The concurrent collector doesn't touch the nursery. */
1761 sgen_nursery_alloc_prepare_for_major ();
1766 /* Pinning depends on this */
1767 sgen_clear_nursery_fragments ();
1769 if (whole_heap_check_before_collection)
1770 sgen_check_whole_heap (TRUE);
1773 time_major_pre_collection_fragment_clear += TV_ELAPSED (atv, btv);
1775 if (!sgen_collection_is_concurrent ())
1776 nursery_section->next_data = sgen_get_nursery_end ();
1777 /* we should also coalesce scanning from sections close to each other
1778 * and deal with pointers outside of the sections later.
1783 sgen_client_pre_collection_checks ();
1785 if (mode != COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT) {
1786 /* Remsets are not useful for a major collection */
1787 remset.clear_cards ();
1790 sgen_process_fin_stage_entries ();
1793 sgen_init_pinning ();
1794 SGEN_LOG (6, "Collecting pinned addresses");
1795 pin_from_roots ((void*)lowest_heap_address, (void*)highest_heap_address, ctx);
1796 if (mode == COPY_OR_MARK_FROM_ROOTS_FINISH_CONCURRENT) {
1797 /* Pin cemented objects that were forced */
1798 sgen_pin_cemented_objects ();
1800 sgen_optimize_pin_queue ();
1801 if (mode == COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT) {
1803 * Cemented objects that are in the pinned list will be marked. When
1804 * marking concurrently we won't mark mod-union cards for these objects.
1805 * Instead they will remain cemented until the next major collection,
1806 * when we will recheck if they are still pinned in the roots.
1808 sgen_cement_force_pinned ();
1811 sgen_client_collecting_major_1 ();
1814 * pin_queue now contains all candidate pointers, sorted and
1815 * uniqued. We must do two passes now to figure out which
1816 * objects are pinned.
1818 * The first is to find within the pin_queue the area for each
1819 * section. This requires that the pin_queue be sorted. We
1820 * also process the LOS objects and pinned chunks here.
1822 * The second, destructive, pass is to reduce the section
1823 * areas to pointers to the actually pinned objects.
1825 SGEN_LOG (6, "Pinning from sections");
1826 /* first pass for the sections */
1827 sgen_find_section_pin_queue_start_end (nursery_section);
1828 /* identify possible pointers to the insize of large objects */
1829 SGEN_LOG (6, "Pinning from large objects");
1830 for (bigobj = los_object_list; bigobj; bigobj = bigobj->next) {
1832 if (sgen_find_optimized_pin_queue_area ((char*)bigobj->data, (char*)bigobj->data + sgen_los_object_size (bigobj), &dummy, &dummy)) {
1833 binary_protocol_pin (bigobj->data, (gpointer)LOAD_VTABLE (bigobj->data), safe_object_get_size (bigobj->data));
1835 if (sgen_los_object_is_pinned (bigobj->data)) {
1836 SGEN_ASSERT (0, mode == COPY_OR_MARK_FROM_ROOTS_FINISH_CONCURRENT, "LOS objects can only be pinned here after concurrent marking.");
1839 sgen_los_pin_object (bigobj->data);
1840 if (SGEN_OBJECT_HAS_REFERENCES (bigobj->data))
1841 GRAY_OBJECT_ENQUEUE_SERIAL (gc_thread_gray_queue, bigobj->data, sgen_obj_get_descriptor ((GCObject*)bigobj->data));
1842 sgen_pin_stats_register_object (bigobj->data, GENERATION_OLD);
1843 SGEN_LOG (6, "Marked large object %p (%s) size: %lu from roots", bigobj->data,
1844 sgen_client_vtable_get_name (SGEN_LOAD_VTABLE (bigobj->data)),
1845 (unsigned long)sgen_los_object_size (bigobj));
1847 sgen_client_pinned_los_object (bigobj->data);
1851 pin_objects_in_nursery (mode == COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT, ctx);
1852 if (check_nursery_objects_pinned && !sgen_minor_collector.is_split)
1853 sgen_check_nursery_objects_pinned (mode != COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT);
1855 major_collector.pin_objects (gc_thread_gray_queue);
1856 if (old_next_pin_slot)
1857 *old_next_pin_slot = sgen_get_pinned_count ();
1860 time_major_pinning += TV_ELAPSED (atv, btv);
1861 SGEN_LOG (2, "Finding pinned pointers: %zd in %lld usecs", sgen_get_pinned_count (), (long long)TV_ELAPSED (atv, btv));
1862 SGEN_LOG (4, "Start scan with %zd pinned objects", sgen_get_pinned_count ());
1864 major_collector.init_to_space ();
1866 SGEN_ASSERT (0, sgen_workers_all_done (), "Why are the workers not done when we start or finish a major collection?");
1867 if (mode == COPY_OR_MARK_FROM_ROOTS_FINISH_CONCURRENT) {
1868 if (object_ops_par != NULL)
1869 sgen_workers_set_num_active_workers (0);
1870 if (sgen_workers_have_idle_work ()) {
1872 * We force the finish of the worker with the new object ops context
1873 * which can also do copying. We need to have finished pinning.
1875 sgen_workers_start_all_workers (object_ops_nopar, object_ops_par, NULL);
1877 sgen_workers_join ();
1881 #ifdef SGEN_DEBUG_INTERNAL_ALLOC
1882 main_gc_thread = mono_native_thread_self ();
1885 sgen_client_collecting_major_2 ();
1888 time_major_scan_pinned += TV_ELAPSED (btv, atv);
1890 sgen_client_collecting_major_3 (&fin_ready_queue, &critical_fin_queue);
1892 enqueue_scan_from_roots_jobs (gc_thread_gray_queue, heap_start, heap_end, object_ops_nopar, FALSE);
1895 time_major_scan_roots += TV_ELAPSED (atv, btv);
1898 * We start the concurrent worker after pinning and after we scanned the roots
1899 * in order to make sure that the worker does not finish before handling all
1902 if (mode == COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT) {
1903 sgen_workers_set_num_active_workers (1);
1904 gray_queue_redirect (gc_thread_gray_queue);
1905 if (precleaning_enabled) {
1906 sgen_workers_start_all_workers (object_ops_nopar, object_ops_par, workers_finish_callback);
1908 sgen_workers_start_all_workers (object_ops_nopar, object_ops_par, NULL);
1912 if (mode == COPY_OR_MARK_FROM_ROOTS_FINISH_CONCURRENT) {
1913 int i, split_count = sgen_workers_get_job_split_count ();
1914 gboolean parallel = object_ops_par != NULL;
1916 /* If we're not parallel we finish the collection on the gc thread */
1918 gray_queue_redirect (gc_thread_gray_queue);
1920 /* Mod union card table */
1921 for (i = 0; i < split_count; i++) {
1922 ParallelScanJob *psj;
1924 psj = (ParallelScanJob*)sgen_thread_pool_job_alloc ("scan mod union cardtable", job_scan_major_mod_union_card_table, sizeof (ParallelScanJob));
1925 psj->scan_job.ops = object_ops_par ? object_ops_par : object_ops_nopar;
1926 psj->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1928 sgen_workers_enqueue_job (&psj->scan_job.job, parallel);
1930 psj = (ParallelScanJob*)sgen_thread_pool_job_alloc ("scan LOS mod union cardtable", job_scan_los_mod_union_card_table, sizeof (ParallelScanJob));
1931 psj->scan_job.ops = object_ops_par ? object_ops_par : object_ops_nopar;
1932 psj->scan_job.gc_thread_gray_queue = gc_thread_gray_queue;
1934 sgen_workers_enqueue_job (&psj->scan_job.job, parallel);
1939 * If we enqueue a job while workers are running we need to sgen_workers_ensure_awake
1940 * in order to make sure that we are running the idle func and draining all worker
1941 * gray queues. The operation of starting workers implies this, so we start them after
1942 * in order to avoid doing this operation twice. The workers will drain the main gray
1943 * stack that contained roots and pinned objects and also scan the mod union card
1946 sgen_workers_start_all_workers (object_ops_nopar, object_ops_par, NULL);
1947 sgen_workers_join ();
1951 sgen_pin_stats_report ();
1953 if (mode == COPY_OR_MARK_FROM_ROOTS_START_CONCURRENT) {
1954 sgen_finish_pinning ();
1956 sgen_pin_stats_reset ();
1958 if (do_concurrent_checks)
1959 sgen_debug_check_nursery_is_clean ();
1964 major_start_collection (SgenGrayQueue *gc_thread_gray_queue, const char *reason, gboolean concurrent, size_t *old_next_pin_slot)
1966 SgenObjectOperations *object_ops_nopar, *object_ops_par = NULL;
1968 binary_protocol_collection_begin (gc_stats.major_gc_count, GENERATION_OLD);
1970 current_collection_generation = GENERATION_OLD;
1972 sgen_workers_assert_gray_queue_is_empty ();
1975 sgen_cement_reset ();
1978 g_assert (major_collector.is_concurrent);
1979 concurrent_collection_in_progress = TRUE;
1981 object_ops_nopar = &major_collector.major_ops_concurrent_start;
1982 if (major_collector.is_parallel)
1983 object_ops_par = &major_collector.major_ops_conc_par_start;
1986 object_ops_nopar = &major_collector.major_ops_serial;
1989 reset_pinned_from_failed_allocation ();
1991 sgen_memgov_major_collection_start (concurrent, reason);
1993 //count_ref_nonref_objs ();
1994 //consistency_check ();
1996 check_scan_starts ();
1999 SGEN_LOG (1, "Start major collection %d", gc_stats.major_gc_count);
2000 gc_stats.major_gc_count ++;
2002 if (major_collector.start_major_collection)
2003 major_collector.start_major_collection ();
2005 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);
2009 major_finish_collection (SgenGrayQueue *gc_thread_gray_queue, const char *reason, gboolean is_overflow, size_t old_next_pin_slot, gboolean forced)
2011 ScannedObjectCounts counts;
2012 SgenObjectOperations *object_ops_nopar;
2013 mword fragment_total;
2019 if (concurrent_collection_in_progress) {
2020 SgenObjectOperations *object_ops_par = NULL;
2022 object_ops_nopar = &major_collector.major_ops_concurrent_finish;
2023 if (major_collector.is_parallel)
2024 object_ops_par = &major_collector.major_ops_conc_par_finish;
2026 major_copy_or_mark_from_roots (gc_thread_gray_queue, NULL, COPY_OR_MARK_FROM_ROOTS_FINISH_CONCURRENT, object_ops_nopar, object_ops_par);
2028 #ifdef SGEN_DEBUG_INTERNAL_ALLOC
2029 main_gc_thread = NULL;
2032 object_ops_nopar = &major_collector.major_ops_serial;
2035 sgen_workers_assert_gray_queue_is_empty ();
2037 finish_gray_stack (GENERATION_OLD, CONTEXT_FROM_OBJECT_OPERATIONS (object_ops_nopar, gc_thread_gray_queue));
2039 time_major_finish_gray_stack += TV_ELAPSED (btv, atv);
2041 SGEN_ASSERT (0, sgen_workers_all_done (), "Can't have workers working after joining");
2043 if (objects_pinned) {
2044 g_assert (!concurrent_collection_in_progress);
2047 * This is slow, but we just OOM'd.
2049 * See comment at `sgen_pin_queue_clear_discarded_entries` for how the pin
2050 * queue is laid out at this point.
2052 sgen_pin_queue_clear_discarded_entries (nursery_section, old_next_pin_slot);
2054 * We need to reestablish all pinned nursery objects in the pin queue
2055 * because they're needed for fragment creation. Unpinning happens by
2056 * walking the whole queue, so it's not necessary to reestablish where major
2057 * heap block pins are - all we care is that they're still in there
2060 sgen_optimize_pin_queue ();
2061 sgen_find_section_pin_queue_start_end (nursery_section);
2065 reset_heap_boundaries ();
2066 sgen_update_heap_boundaries ((mword)sgen_get_nursery_start (), (mword)sgen_get_nursery_end ());
2068 /* walk the pin_queue, build up the fragment list of free memory, unmark
2069 * pinned objects as we go, memzero() the empty fragments so they are ready for the
2072 fragment_total = sgen_build_nursery_fragments (nursery_section, NULL);
2073 if (!fragment_total)
2075 SGEN_LOG (4, "Free space in nursery after major %ld", (long)fragment_total);
2077 if (do_concurrent_checks && concurrent_collection_in_progress)
2078 sgen_debug_check_nursery_is_clean ();
2080 /* prepare the pin queue for the next collection */
2081 sgen_finish_pinning ();
2083 /* Clear TLABs for all threads */
2084 sgen_clear_tlabs ();
2086 sgen_pin_stats_reset ();
2088 sgen_cement_clear_below_threshold ();
2090 if (check_mark_bits_after_major_collection)
2091 sgen_check_heap_marked (concurrent_collection_in_progress);
2094 time_major_fragment_creation += TV_ELAPSED (atv, btv);
2096 binary_protocol_sweep_begin (GENERATION_OLD, !major_collector.sweeps_lazily);
2097 sgen_memgov_major_pre_sweep ();
2100 time_major_free_bigobjs += TV_ELAPSED (btv, atv);
2105 time_major_los_sweep += TV_ELAPSED (atv, btv);
2107 major_collector.sweep ();
2109 binary_protocol_sweep_end (GENERATION_OLD, !major_collector.sweeps_lazily);
2112 time_major_sweep += TV_ELAPSED (btv, atv);
2114 sgen_debug_dump_heap ("major", gc_stats.major_gc_count - 1, reason);
2116 if (sgen_have_pending_finalizers ()) {
2117 SGEN_LOG (4, "Finalizer-thread wakeup");
2118 sgen_client_finalize_notify ();
2121 sgen_memgov_major_collection_end (forced, concurrent_collection_in_progress, reason, is_overflow);
2122 current_collection_generation = -1;
2124 memset (&counts, 0, sizeof (ScannedObjectCounts));
2125 major_collector.finish_major_collection (&counts);
2127 sgen_workers_assert_gray_queue_is_empty ();
2129 SGEN_ASSERT (0, sgen_workers_all_done (), "Can't have workers working after major collection has finished");
2130 if (concurrent_collection_in_progress)
2131 concurrent_collection_in_progress = FALSE;
2133 check_scan_starts ();
2135 binary_protocol_flush_buffers (FALSE);
2137 //consistency_check ();
2139 binary_protocol_collection_end (gc_stats.major_gc_count - 1, GENERATION_OLD, counts.num_scanned_objects, counts.num_unique_scanned_objects);
2143 major_do_collection (const char *reason, gboolean is_overflow, gboolean forced)
2145 TV_DECLARE (time_start);
2146 TV_DECLARE (time_end);
2147 size_t old_next_pin_slot;
2148 SgenGrayQueue gc_thread_gray_queue;
2150 if (disable_major_collections)
2153 if (major_collector.get_and_reset_num_major_objects_marked) {
2154 long long num_marked = major_collector.get_and_reset_num_major_objects_marked ();
2155 g_assert (!num_marked);
2158 /* world must be stopped already */
2159 TV_GETTIME (time_start);
2161 init_gray_queue (&gc_thread_gray_queue, FALSE);
2162 major_start_collection (&gc_thread_gray_queue, reason, FALSE, &old_next_pin_slot);
2163 major_finish_collection (&gc_thread_gray_queue, reason, is_overflow, old_next_pin_slot, forced);
2164 sgen_gray_object_queue_dispose (&gc_thread_gray_queue);
2166 TV_GETTIME (time_end);
2167 gc_stats.major_gc_time += TV_ELAPSED (time_start, time_end);
2169 /* FIXME: also report this to the user, preferably in gc-end. */
2170 if (major_collector.get_and_reset_num_major_objects_marked)
2171 major_collector.get_and_reset_num_major_objects_marked ();
2173 return bytes_pinned_from_failed_allocation > 0;
2177 major_start_concurrent_collection (const char *reason)
2179 TV_DECLARE (time_start);
2180 TV_DECLARE (time_end);
2181 long long num_objects_marked;
2182 SgenGrayQueue gc_thread_gray_queue;
2184 if (disable_major_collections)
2187 TV_GETTIME (time_start);
2188 SGEN_TV_GETTIME (time_major_conc_collection_start);
2190 num_objects_marked = major_collector.get_and_reset_num_major_objects_marked ();
2191 g_assert (num_objects_marked == 0);
2193 binary_protocol_concurrent_start ();
2195 init_gray_queue (&gc_thread_gray_queue, TRUE);
2196 // FIXME: store reason and pass it when finishing
2197 major_start_collection (&gc_thread_gray_queue, reason, TRUE, NULL);
2198 sgen_gray_object_queue_dispose (&gc_thread_gray_queue);
2200 num_objects_marked = major_collector.get_and_reset_num_major_objects_marked ();
2202 TV_GETTIME (time_end);
2203 gc_stats.major_gc_time += TV_ELAPSED (time_start, time_end);
2205 current_collection_generation = -1;
2209 * Returns whether the major collection has finished.
2212 major_should_finish_concurrent_collection (void)
2214 return sgen_workers_all_done ();
2218 major_update_concurrent_collection (void)
2220 TV_DECLARE (total_start);
2221 TV_DECLARE (total_end);
2223 TV_GETTIME (total_start);
2225 binary_protocol_concurrent_update ();
2227 major_collector.update_cardtable_mod_union ();
2228 sgen_los_update_cardtable_mod_union ();
2230 TV_GETTIME (total_end);
2231 gc_stats.major_gc_time += TV_ELAPSED (total_start, total_end);
2235 major_finish_concurrent_collection (gboolean forced)
2237 SgenGrayQueue gc_thread_gray_queue;
2238 TV_DECLARE (total_start);
2239 TV_DECLARE (total_end);
2241 TV_GETTIME (total_start);
2243 binary_protocol_concurrent_finish ();
2246 * We need to stop all workers since we're updating the cardtable below.
2247 * The workers will be resumed with a finishing pause context to avoid
2248 * additional cardtable and object scanning.
2250 sgen_workers_stop_all_workers ();
2252 SGEN_TV_GETTIME (time_major_conc_collection_end);
2253 gc_stats.major_gc_time_concurrent += SGEN_TV_ELAPSED (time_major_conc_collection_start, time_major_conc_collection_end);
2255 major_collector.update_cardtable_mod_union ();
2256 sgen_los_update_cardtable_mod_union ();
2258 if (mod_union_consistency_check)
2259 sgen_check_mod_union_consistency ();
2261 current_collection_generation = GENERATION_OLD;
2262 sgen_cement_reset ();
2263 init_gray_queue (&gc_thread_gray_queue, FALSE);
2264 major_finish_collection (&gc_thread_gray_queue, "finishing", FALSE, -1, forced);
2265 sgen_gray_object_queue_dispose (&gc_thread_gray_queue);
2267 TV_GETTIME (total_end);
2268 gc_stats.major_gc_time += TV_ELAPSED (total_start, total_end);
2270 current_collection_generation = -1;
2274 * Ensure an allocation request for @size will succeed by freeing enough memory.
2276 * LOCKING: The GC lock MUST be held.
2279 sgen_ensure_free_space (size_t size, int generation)
2281 int generation_to_collect = -1;
2282 const char *reason = NULL;
2284 if (generation == GENERATION_OLD) {
2285 if (sgen_need_major_collection (size)) {
2286 reason = "LOS overflow";
2287 generation_to_collect = GENERATION_OLD;
2290 if (degraded_mode) {
2291 if (sgen_need_major_collection (size)) {
2292 reason = "Degraded mode overflow";
2293 generation_to_collect = GENERATION_OLD;
2295 } else if (sgen_need_major_collection (size)) {
2296 reason = concurrent_collection_in_progress ? "Forced finish concurrent collection" : "Minor allowance";
2297 generation_to_collect = GENERATION_OLD;
2299 generation_to_collect = GENERATION_NURSERY;
2300 reason = "Nursery full";
2304 if (generation_to_collect == -1) {
2305 if (concurrent_collection_in_progress && sgen_workers_all_done ()) {
2306 generation_to_collect = GENERATION_OLD;
2307 reason = "Finish concurrent collection";
2311 if (generation_to_collect == -1)
2313 sgen_perform_collection (size, generation_to_collect, reason, FALSE, TRUE);
2317 * LOCKING: Assumes the GC lock is held.
2320 sgen_perform_collection (size_t requested_size, int generation_to_collect, const char *reason, gboolean wait_to_finish, gboolean stw)
2322 TV_DECLARE (gc_total_start);
2323 TV_DECLARE (gc_total_end);
2324 int overflow_generation_to_collect = -1;
2325 int oldest_generation_collected = generation_to_collect;
2326 const char *overflow_reason = NULL;
2327 gboolean finish_concurrent = concurrent_collection_in_progress && (major_should_finish_concurrent_collection () || generation_to_collect == GENERATION_OLD);
2329 binary_protocol_collection_requested (generation_to_collect, requested_size, wait_to_finish ? 1 : 0);
2331 SGEN_ASSERT (0, generation_to_collect == GENERATION_NURSERY || generation_to_collect == GENERATION_OLD, "What generation is this?");
2334 sgen_stop_world (generation_to_collect);
2336 SGEN_ASSERT (0, sgen_is_world_stopped (), "We can only collect if the world is stopped");
2339 TV_GETTIME (gc_total_start);
2341 // FIXME: extract overflow reason
2342 // FIXME: minor overflow for concurrent case
2343 if (generation_to_collect == GENERATION_NURSERY && !finish_concurrent) {
2344 if (concurrent_collection_in_progress)
2345 major_update_concurrent_collection ();
2347 if (collect_nursery (reason, FALSE, NULL) && !concurrent_collection_in_progress) {
2348 overflow_generation_to_collect = GENERATION_OLD;
2349 overflow_reason = "Minor overflow";
2351 } else if (finish_concurrent) {
2352 major_finish_concurrent_collection (wait_to_finish);
2353 oldest_generation_collected = GENERATION_OLD;
2355 SGEN_ASSERT (0, generation_to_collect == GENERATION_OLD, "We should have handled nursery collections above");
2356 if (major_collector.is_concurrent && !wait_to_finish) {
2357 collect_nursery ("Concurrent start", FALSE, NULL);
2358 major_start_concurrent_collection (reason);
2359 oldest_generation_collected = GENERATION_NURSERY;
2360 } else if (major_do_collection (reason, FALSE, wait_to_finish)) {
2361 overflow_generation_to_collect = GENERATION_NURSERY;
2362 overflow_reason = "Excessive pinning";
2366 if (overflow_generation_to_collect != -1) {
2367 SGEN_ASSERT (0, !concurrent_collection_in_progress, "We don't yet support overflow collections with the concurrent collector");
2370 * We need to do an overflow collection, either because we ran out of memory
2371 * or the nursery is fully pinned.
2374 if (overflow_generation_to_collect == GENERATION_NURSERY)
2375 collect_nursery (overflow_reason, TRUE, NULL);
2377 major_do_collection (overflow_reason, TRUE, wait_to_finish);
2379 oldest_generation_collected = MAX (oldest_generation_collected, overflow_generation_to_collect);
2382 SGEN_LOG (2, "Heap size: %lu, LOS size: %lu", (unsigned long)sgen_gc_get_total_heap_allocation (), (unsigned long)los_memory_usage);
2384 /* this also sets the proper pointers for the next allocation */
2385 if (generation_to_collect == GENERATION_NURSERY && !sgen_can_alloc_size (requested_size)) {
2386 /* TypeBuilder and MonoMethod are killing mcs with fragmentation */
2387 SGEN_LOG (1, "nursery collection didn't find enough room for %zd alloc (%zd pinned)", requested_size, sgen_get_pinned_count ());
2388 sgen_dump_pin_queue ();
2392 TV_GETTIME (gc_total_end);
2393 time_max = MAX (time_max, TV_ELAPSED (gc_total_start, gc_total_end));
2396 sgen_restart_world (oldest_generation_collected);
2400 * ######################################################################
2401 * ######## Memory allocation from the OS
2402 * ######################################################################
2403 * This section of code deals with getting memory from the OS and
2404 * allocating memory for GC-internal data structures.
2405 * Internal memory can be handled with a freelist for small objects.
2411 G_GNUC_UNUSED static void
2412 report_internal_mem_usage (void)
2414 printf ("Internal memory usage:\n");
2415 sgen_report_internal_mem_usage ();
2416 printf ("Pinned memory usage:\n");
2417 major_collector.report_pinned_memory_usage ();
2421 * ######################################################################
2422 * ######## Finalization support
2423 * ######################################################################
2427 * If the object has been forwarded it means it's still referenced from a root.
2428 * If it is pinned it's still alive as well.
2429 * A LOS object is only alive if we have pinned it.
2430 * Return TRUE if @obj is ready to be finalized.
2432 static inline gboolean
2433 sgen_is_object_alive (GCObject *object)
2435 if (ptr_in_nursery (object))
2436 return sgen_nursery_is_object_alive (object);
2438 return sgen_major_is_object_alive (object);
2442 * This function returns true if @object is either alive and belongs to the
2443 * current collection - major collections are full heap, so old gen objects
2444 * are never alive during a minor collection.
2447 sgen_is_object_alive_and_on_current_collection (GCObject *object)
2449 if (ptr_in_nursery (object))
2450 return sgen_nursery_is_object_alive (object);
2452 if (current_collection_generation == GENERATION_NURSERY)
2455 return sgen_major_is_object_alive (object);
2460 sgen_gc_is_object_ready_for_finalization (GCObject *object)
2462 return !sgen_is_object_alive (object);
2466 sgen_queue_finalization_entry (GCObject *obj)
2468 gboolean critical = sgen_client_object_has_critical_finalizer (obj);
2470 sgen_pointer_queue_add (critical ? &critical_fin_queue : &fin_ready_queue, obj);
2472 sgen_client_object_queued_for_finalization (obj);
2476 sgen_object_is_live (GCObject *obj)
2478 return sgen_is_object_alive_and_on_current_collection (obj);
2482 * `System.GC.WaitForPendingFinalizers` first checks `sgen_have_pending_finalizers()` to
2483 * determine whether it can exit quickly. The latter must therefore only return FALSE if
2484 * all finalizers have really finished running.
2486 * `sgen_gc_invoke_finalizers()` first dequeues a finalizable object, and then finalizes it.
2487 * This means that just checking whether the queues are empty leaves the possibility that an
2488 * object might have been dequeued but not yet finalized. That's why we need the additional
2489 * flag `pending_unqueued_finalizer`.
2492 static volatile gboolean pending_unqueued_finalizer = FALSE;
2493 volatile gboolean sgen_suspend_finalizers = FALSE;
2496 sgen_set_suspend_finalizers (void)
2498 sgen_suspend_finalizers = TRUE;
2502 sgen_gc_invoke_finalizers (void)
2506 g_assert (!pending_unqueued_finalizer);
2508 /* FIXME: batch to reduce lock contention */
2509 while (sgen_have_pending_finalizers ()) {
2515 * We need to set `pending_unqueued_finalizer` before dequeing the
2516 * finalizable object.
2518 if (!sgen_pointer_queue_is_empty (&fin_ready_queue)) {
2519 pending_unqueued_finalizer = TRUE;
2520 mono_memory_write_barrier ();
2521 obj = (GCObject *)sgen_pointer_queue_pop (&fin_ready_queue);
2522 } else if (!sgen_pointer_queue_is_empty (&critical_fin_queue)) {
2523 pending_unqueued_finalizer = TRUE;
2524 mono_memory_write_barrier ();
2525 obj = (GCObject *)sgen_pointer_queue_pop (&critical_fin_queue);
2531 SGEN_LOG (7, "Finalizing object %p (%s)", obj, sgen_client_vtable_get_name (SGEN_LOAD_VTABLE (obj)));
2539 /* the object is on the stack so it is pinned */
2540 /*g_print ("Calling finalizer for object: %p (%s)\n", obj, sgen_client_object_safe_name (obj));*/
2541 sgen_client_run_finalize (obj);
2544 if (pending_unqueued_finalizer) {
2545 mono_memory_write_barrier ();
2546 pending_unqueued_finalizer = FALSE;
2553 sgen_have_pending_finalizers (void)
2555 if (sgen_suspend_finalizers)
2557 return pending_unqueued_finalizer || !sgen_pointer_queue_is_empty (&fin_ready_queue) || !sgen_pointer_queue_is_empty (&critical_fin_queue);
2561 * ######################################################################
2562 * ######## registered roots support
2563 * ######################################################################
2567 * We do not coalesce roots.
2570 sgen_register_root (char *start, size_t size, SgenDescriptor descr, int root_type, int source, const char *msg)
2572 RootRecord new_root;
2575 for (i = 0; i < ROOT_TYPE_NUM; ++i) {
2576 RootRecord *root = (RootRecord *)sgen_hash_table_lookup (&roots_hash [i], start);
2577 /* we allow changing the size and the descriptor (for thread statics etc) */
2579 size_t old_size = root->end_root - start;
2580 root->end_root = start + size;
2581 SGEN_ASSERT (0, !!root->root_desc == !!descr, "Can't change whether a root is precise or conservative.");
2582 SGEN_ASSERT (0, root->source == source, "Can't change a root's source identifier.");
2583 SGEN_ASSERT (0, !!root->msg == !!msg, "Can't change a root's message.");
2584 root->root_desc = descr;
2586 roots_size -= old_size;
2592 new_root.end_root = start + size;
2593 new_root.root_desc = descr;
2594 new_root.source = source;
2597 sgen_hash_table_replace (&roots_hash [root_type], start, &new_root, NULL);
2600 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);
2607 sgen_deregister_root (char* addr)
2613 for (root_type = 0; root_type < ROOT_TYPE_NUM; ++root_type) {
2614 if (sgen_hash_table_remove (&roots_hash [root_type], addr, &root))
2615 roots_size -= (root.end_root - addr);
2621 sgen_wbroots_iterate_live_block_ranges (sgen_cardtable_block_callback cb)
2625 SGEN_HASH_TABLE_FOREACH (&roots_hash [ROOT_TYPE_WBARRIER], void **, start_root, RootRecord *, root) {
2626 cb ((mword)start_root, (mword)root->end_root - (mword)start_root);
2627 } SGEN_HASH_TABLE_FOREACH_END;
2630 /* Root equivalent of sgen_client_cardtable_scan_object */
2632 sgen_wbroot_scan_card_table (void** start_root, mword size, ScanCopyContext ctx)
2634 ScanPtrFieldFunc scan_field_func = ctx.ops->scan_ptr_field;
2635 guint8 *card_data = sgen_card_table_get_card_scan_address ((mword)start_root);
2636 guint8 *card_base = card_data;
2637 mword card_count = sgen_card_table_number_of_cards_in_range ((mword)start_root, size);
2638 guint8 *card_data_end = card_data + card_count;
2639 mword extra_idx = 0;
2640 char *obj_start = sgen_card_table_align_pointer (start_root);
2641 char *obj_end = (char*)start_root + size;
2642 #ifdef SGEN_HAVE_OVERLAPPING_CARDS
2643 guint8 *overflow_scan_end = NULL;
2646 #ifdef SGEN_HAVE_OVERLAPPING_CARDS
2647 /*Check for overflow and if so, setup to scan in two steps*/
2648 if (card_data_end >= SGEN_SHADOW_CARDTABLE_END) {
2649 overflow_scan_end = sgen_shadow_cardtable + (card_data_end - SGEN_SHADOW_CARDTABLE_END);
2650 card_data_end = SGEN_SHADOW_CARDTABLE_END;
2656 card_data = sgen_find_next_card (card_data, card_data_end);
2658 for (; card_data < card_data_end; card_data = sgen_find_next_card (card_data + 1, card_data_end)) {
2659 size_t idx = (card_data - card_base) + extra_idx;
2660 char *start = (char*)(obj_start + idx * CARD_SIZE_IN_BYTES);
2661 char *card_end = start + CARD_SIZE_IN_BYTES;
2662 char *elem = start, *first_elem = start;
2665 * Don't clean first and last card on 32bit systems since they
2666 * may also be part from other roots.
2668 if (card_data != card_base && card_data != (card_data_end - 1))
2669 sgen_card_table_prepare_card_for_scanning (card_data);
2671 card_end = MIN (card_end, obj_end);
2673 if (elem < (char*)start_root)
2674 first_elem = elem = (char*)start_root;
2676 for (; elem < card_end; elem += SIZEOF_VOID_P) {
2677 if (*(GCObject**)elem)
2678 scan_field_func (NULL, (GCObject**)elem, ctx.queue);
2681 binary_protocol_card_scan (first_elem, elem - first_elem);
2684 #ifdef SGEN_HAVE_OVERLAPPING_CARDS
2685 if (overflow_scan_end) {
2686 extra_idx = card_data - card_base;
2687 card_base = card_data = sgen_shadow_cardtable;
2688 card_data_end = overflow_scan_end;
2689 overflow_scan_end = NULL;
2696 sgen_wbroots_scan_card_table (ScanCopyContext ctx)
2701 SGEN_HASH_TABLE_FOREACH (&roots_hash [ROOT_TYPE_WBARRIER], void **, start_root, RootRecord *, root) {
2702 SGEN_ASSERT (0, (root->root_desc & ROOT_DESC_TYPE_MASK) == ROOT_DESC_VECTOR, "Unsupported root type");
2704 sgen_wbroot_scan_card_table (start_root, (mword)root->end_root - (mword)start_root, ctx);
2705 } SGEN_HASH_TABLE_FOREACH_END;
2709 * ######################################################################
2710 * ######## Thread handling (stop/start code)
2711 * ######################################################################
2715 sgen_get_current_collection_generation (void)
2717 return current_collection_generation;
2721 sgen_thread_register (SgenThreadInfo* info, void *stack_bottom_fallback)
2723 info->tlab_start = info->tlab_next = info->tlab_temp_end = info->tlab_real_end = NULL;
2725 sgen_client_thread_register (info, stack_bottom_fallback);
2731 sgen_thread_unregister (SgenThreadInfo *p)
2733 sgen_client_thread_unregister (p);
2737 * ######################################################################
2738 * ######## Write barriers
2739 * ######################################################################
2743 * Note: the write barriers first do the needed GC work and then do the actual store:
2744 * this way the value is visible to the conservative GC scan after the write barrier
2745 * itself. If a GC interrupts the barrier in the middle, value will be kept alive by
2746 * the conservative scan, otherwise by the remembered set scan.
2750 * mono_gc_wbarrier_arrayref_copy:
2753 mono_gc_wbarrier_arrayref_copy (gpointer dest_ptr, gpointer src_ptr, int count)
2755 HEAVY_STAT (++stat_wbarrier_arrayref_copy);
2756 /*This check can be done without taking a lock since dest_ptr array is pinned*/
2757 if (ptr_in_nursery (dest_ptr) || count <= 0) {
2758 mono_gc_memmove_aligned (dest_ptr, src_ptr, count * sizeof (gpointer));
2762 #ifdef SGEN_HEAVY_BINARY_PROTOCOL
2763 if (binary_protocol_is_heavy_enabled ()) {
2765 for (i = 0; i < count; ++i) {
2766 gpointer dest = (gpointer*)dest_ptr + i;
2767 gpointer obj = *((gpointer*)src_ptr + i);
2769 binary_protocol_wbarrier (dest, obj, (gpointer)LOAD_VTABLE (obj));
2774 remset.wbarrier_arrayref_copy (dest_ptr, src_ptr, count);
2778 * mono_gc_wbarrier_generic_nostore:
2781 mono_gc_wbarrier_generic_nostore (gpointer ptr)
2785 HEAVY_STAT (++stat_wbarrier_generic_store);
2787 sgen_client_wbarrier_generic_nostore_check (ptr);
2789 obj = *(gpointer*)ptr;
2791 binary_protocol_wbarrier (ptr, obj, (gpointer)LOAD_VTABLE (obj));
2794 * We need to record old->old pointer locations for the
2795 * concurrent collector.
2797 if (!ptr_in_nursery (obj) && !concurrent_collection_in_progress) {
2798 SGEN_LOG (8, "Skipping remset at %p", ptr);
2802 SGEN_LOG (8, "Adding remset at %p", ptr);
2804 remset.wbarrier_generic_nostore (ptr);
2808 * mono_gc_wbarrier_generic_store:
2811 mono_gc_wbarrier_generic_store (gpointer ptr, GCObject* value)
2813 SGEN_LOG (8, "Wbarrier store at %p to %p (%s)", ptr, value, value ? sgen_client_vtable_get_name (SGEN_LOAD_VTABLE (value)) : "null");
2814 SGEN_UPDATE_REFERENCE_ALLOW_NULL (ptr, value);
2815 if (ptr_in_nursery (value) || concurrent_collection_in_progress)
2816 mono_gc_wbarrier_generic_nostore (ptr);
2817 sgen_dummy_use (value);
2821 * mono_gc_wbarrier_generic_store_atomic:
2822 * Same as \c mono_gc_wbarrier_generic_store but performs the store
2823 * as an atomic operation with release semantics.
2826 mono_gc_wbarrier_generic_store_atomic (gpointer ptr, GCObject *value)
2828 HEAVY_STAT (++stat_wbarrier_generic_store_atomic);
2830 SGEN_LOG (8, "Wbarrier atomic store at %p to %p (%s)", ptr, value, value ? sgen_client_vtable_get_name (SGEN_LOAD_VTABLE (value)) : "null");
2832 InterlockedWritePointer ((volatile gpointer *)ptr, value);
2834 if (ptr_in_nursery (value) || concurrent_collection_in_progress)
2835 mono_gc_wbarrier_generic_nostore (ptr);
2837 sgen_dummy_use (value);
2841 sgen_wbarrier_value_copy_bitmap (gpointer _dest, gpointer _src, int size, unsigned bitmap)
2843 GCObject **dest = (GCObject **)_dest;
2844 GCObject **src = (GCObject **)_src;
2848 mono_gc_wbarrier_generic_store (dest, *src);
2853 size -= SIZEOF_VOID_P;
2859 * ######################################################################
2860 * ######## Other mono public interface functions.
2861 * ######################################################################
2865 sgen_gc_collect (int generation)
2870 sgen_perform_collection (0, generation, "user request", TRUE, TRUE);
2875 sgen_gc_collection_count (int generation)
2877 if (generation == 0)
2878 return gc_stats.minor_gc_count;
2879 return gc_stats.major_gc_count;
2883 sgen_gc_get_used_size (void)
2887 tot = los_memory_usage;
2888 tot += nursery_section->next_data - nursery_section->data;
2889 tot += major_collector.get_used_size ();
2890 /* FIXME: account for pinned objects */
2896 sgen_env_var_error (const char *env_var, const char *fallback, const char *description_format, ...)
2900 va_start (ap, description_format);
2902 fprintf (stderr, "Warning: In environment variable `%s': ", env_var);
2903 vfprintf (stderr, description_format, ap);
2905 fprintf (stderr, " - %s", fallback);
2906 fprintf (stderr, "\n");
2912 parse_double_in_interval (const char *env_var, const char *opt_name, const char *opt, double min, double max, double *result)
2915 double val = strtod (opt, &endptr);
2916 if (endptr == opt) {
2917 sgen_env_var_error (env_var, "Using default value.", "`%s` must be a number.", opt_name);
2920 else if (val < min || val > max) {
2921 sgen_env_var_error (env_var, "Using default value.", "`%s` must be between %.2f - %.2f.", opt_name, min, max);
2933 char *major_collector_opt = NULL;
2934 char *minor_collector_opt = NULL;
2935 char *params_opts = NULL;
2936 char *debug_opts = NULL;
2937 size_t max_heap = 0;
2938 size_t soft_limit = 0;
2940 gboolean debug_print_allowance = FALSE;
2941 double allowance_ratio = 0, save_target = 0;
2942 gboolean cement_enabled = TRUE;
2945 result = InterlockedCompareExchange (&gc_initialized, -1, 0);
2948 /* already inited */
2951 /* being inited by another thread */
2952 mono_thread_info_usleep (1000);
2955 /* we will init it */
2958 g_assert_not_reached ();
2960 } while (result != 0);
2962 SGEN_TV_GETTIME (sgen_init_timestamp);
2964 #ifdef SGEN_WITHOUT_MONO
2965 mono_thread_smr_init ();
2968 mono_coop_mutex_init (&gc_mutex);
2970 gc_debug_file = stderr;
2972 mono_coop_mutex_init (&sgen_interruption_mutex);
2974 if ((env = g_getenv (MONO_GC_PARAMS_NAME)) || gc_params_options) {
2975 params_opts = g_strdup_printf ("%s,%s", gc_params_options ? gc_params_options : "", env ? env : "");
2979 opts = g_strsplit (params_opts, ",", -1);
2980 for (ptr = opts; *ptr; ++ptr) {
2982 if (g_str_has_prefix (opt, "major=")) {
2983 opt = strchr (opt, '=') + 1;
2984 major_collector_opt = g_strdup (opt);
2985 } else if (g_str_has_prefix (opt, "minor=")) {
2986 opt = strchr (opt, '=') + 1;
2987 minor_collector_opt = g_strdup (opt);
2995 sgen_init_internal_allocator ();
2996 sgen_init_nursery_allocator ();
2997 sgen_init_fin_weak_hash ();
2998 sgen_init_hash_table ();
2999 sgen_init_descriptors ();
3000 sgen_init_gray_queues ();
3001 sgen_init_allocator ();
3002 sgen_init_gchandles ();
3004 sgen_register_fixed_internal_mem_type (INTERNAL_MEM_SECTION, SGEN_SIZEOF_GC_MEM_SECTION);
3005 sgen_register_fixed_internal_mem_type (INTERNAL_MEM_GRAY_QUEUE, sizeof (GrayQueueSection));
3007 sgen_client_init ();
3009 if (!minor_collector_opt) {
3010 sgen_simple_nursery_init (&sgen_minor_collector, FALSE);
3012 if (!strcmp (minor_collector_opt, "simple")) {
3014 sgen_simple_nursery_init (&sgen_minor_collector, FALSE);
3015 } else if (!strcmp (minor_collector_opt, "simple-par")) {
3016 sgen_simple_nursery_init (&sgen_minor_collector, TRUE);
3017 } else if (!strcmp (minor_collector_opt, "split")) {
3018 sgen_split_nursery_init (&sgen_minor_collector);
3020 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Using `simple` instead.", "Unknown minor collector `%s'.", minor_collector_opt);
3021 goto use_simple_nursery;
3025 if (!major_collector_opt) {
3027 DEFAULT_MAJOR_INIT (&major_collector);
3028 } else if (!strcmp (major_collector_opt, "marksweep")) {
3029 sgen_marksweep_init (&major_collector);
3030 } else if (!strcmp (major_collector_opt, "marksweep-conc")) {
3031 sgen_marksweep_conc_init (&major_collector);
3032 } else if (!strcmp (major_collector_opt, "marksweep-conc-par")) {
3033 sgen_marksweep_conc_par_init (&major_collector);
3035 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Using `" DEFAULT_MAJOR_NAME "` instead.", "Unknown major collector `%s'.", major_collector_opt);
3036 goto use_default_major;
3039 sgen_nursery_size = DEFAULT_NURSERY_SIZE;
3042 gboolean usage_printed = FALSE;
3044 for (ptr = opts; *ptr; ++ptr) {
3046 if (!strcmp (opt, ""))
3048 if (g_str_has_prefix (opt, "major="))
3050 if (g_str_has_prefix (opt, "minor="))
3052 if (g_str_has_prefix (opt, "max-heap-size=")) {
3053 size_t page_size = mono_pagesize ();
3054 size_t max_heap_candidate = 0;
3055 opt = strchr (opt, '=') + 1;
3056 if (*opt && mono_gc_parse_environment_string_extract_number (opt, &max_heap_candidate)) {
3057 max_heap = (max_heap_candidate + page_size - 1) & ~(size_t)(page_size - 1);
3058 if (max_heap != max_heap_candidate)
3059 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Rounding up.", "`max-heap-size` size must be a multiple of %d.", page_size);
3061 sgen_env_var_error (MONO_GC_PARAMS_NAME, NULL, "`max-heap-size` must be an integer.");
3065 if (g_str_has_prefix (opt, "soft-heap-limit=")) {
3066 opt = strchr (opt, '=') + 1;
3067 if (*opt && mono_gc_parse_environment_string_extract_number (opt, &soft_limit)) {
3068 if (soft_limit <= 0) {
3069 sgen_env_var_error (MONO_GC_PARAMS_NAME, NULL, "`soft-heap-limit` must be positive.");
3073 sgen_env_var_error (MONO_GC_PARAMS_NAME, NULL, "`soft-heap-limit` must be an integer.");
3079 if (g_str_has_prefix (opt, "nursery-size=")) {
3081 opt = strchr (opt, '=') + 1;
3082 if (*opt && mono_gc_parse_environment_string_extract_number (opt, &val)) {
3083 if ((val & (val - 1))) {
3084 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Using default value.", "`nursery-size` must be a power of two.");
3088 if (val < SGEN_MAX_NURSERY_WASTE) {
3089 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Using default value.",
3090 "`nursery-size` must be at least %d bytes.", SGEN_MAX_NURSERY_WASTE);
3094 sgen_nursery_size = val;
3095 sgen_nursery_bits = 0;
3096 while (ONE_P << (++ sgen_nursery_bits) != sgen_nursery_size)
3099 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Using default value.", "`nursery-size` must be an integer.");
3105 if (g_str_has_prefix (opt, "save-target-ratio=")) {
3107 opt = strchr (opt, '=') + 1;
3108 if (parse_double_in_interval (MONO_GC_PARAMS_NAME, "save-target-ratio", opt,
3109 SGEN_MIN_SAVE_TARGET_RATIO, SGEN_MAX_SAVE_TARGET_RATIO, &val)) {
3114 if (g_str_has_prefix (opt, "default-allowance-ratio=")) {
3116 opt = strchr (opt, '=') + 1;
3117 if (parse_double_in_interval (MONO_GC_PARAMS_NAME, "default-allowance-ratio", opt,
3118 SGEN_MIN_ALLOWANCE_NURSERY_SIZE_RATIO, SGEN_MAX_ALLOWANCE_NURSERY_SIZE_RATIO, &val)) {
3119 allowance_ratio = val;
3124 if (!strcmp (opt, "cementing")) {
3125 cement_enabled = TRUE;
3128 if (!strcmp (opt, "no-cementing")) {
3129 cement_enabled = FALSE;
3133 if (!strcmp (opt, "precleaning")) {
3134 precleaning_enabled = TRUE;
3137 if (!strcmp (opt, "no-precleaning")) {
3138 precleaning_enabled = FALSE;
3142 if (major_collector.handle_gc_param && major_collector.handle_gc_param (opt))
3145 if (sgen_minor_collector.handle_gc_param && sgen_minor_collector.handle_gc_param (opt))
3148 if (sgen_client_handle_gc_param (opt))
3151 sgen_env_var_error (MONO_GC_PARAMS_NAME, "Ignoring.", "Unknown option `%s`.", opt);
3156 fprintf (stderr, "\n%s must be a comma-delimited list of one or more of the following:\n", MONO_GC_PARAMS_NAME);
3157 fprintf (stderr, " max-heap-size=N (where N is an integer, possibly with a k, m or a g suffix)\n");
3158 fprintf (stderr, " soft-heap-limit=n (where N is an integer, possibly with a k, m or a g suffix)\n");
3159 fprintf (stderr, " nursery-size=N (where N is an integer, possibly with a k, m or a g suffix)\n");
3160 fprintf (stderr, " major=COLLECTOR (where COLLECTOR is `marksweep', `marksweep-conc', `marksweep-par')\n");
3161 fprintf (stderr, " minor=COLLECTOR (where COLLECTOR is `simple' or `split')\n");
3162 fprintf (stderr, " wbarrier=WBARRIER (where WBARRIER is `remset' or `cardtable')\n");
3163 fprintf (stderr, " [no-]cementing\n");
3164 if (major_collector.print_gc_param_usage)
3165 major_collector.print_gc_param_usage ();
3166 if (sgen_minor_collector.print_gc_param_usage)
3167 sgen_minor_collector.print_gc_param_usage ();
3168 sgen_client_print_gc_params_usage ();
3169 fprintf (stderr, " Experimental options:\n");
3170 fprintf (stderr, " save-target-ratio=R (where R must be between %.2f - %.2f).\n", SGEN_MIN_SAVE_TARGET_RATIO, SGEN_MAX_SAVE_TARGET_RATIO);
3171 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);
3172 fprintf (stderr, "\n");
3174 usage_printed = TRUE;
3179 if (major_collector_opt)
3180 g_free (major_collector_opt);
3182 if (minor_collector_opt)
3183 g_free (minor_collector_opt);
3186 g_free (params_opts);
3190 sgen_pinning_init ();
3191 sgen_cement_init (cement_enabled);
3193 if ((env = g_getenv (MONO_GC_DEBUG_NAME)) || gc_debug_options) {
3194 debug_opts = g_strdup_printf ("%s,%s", gc_debug_options ? gc_debug_options : "", env ? env : "");
3198 gboolean usage_printed = FALSE;
3200 opts = g_strsplit (debug_opts, ",", -1);
3201 for (ptr = opts; ptr && *ptr; ptr ++) {
3203 if (!strcmp (opt, ""))
3205 if (opt [0] >= '0' && opt [0] <= '9') {
3206 gc_debug_level = atoi (opt);
3211 char *rf = g_strdup_printf ("%s.%d", opt, mono_process_current_pid ());
3212 gc_debug_file = fopen (rf, "wb");
3214 gc_debug_file = stderr;
3217 } else if (!strcmp (opt, "print-allowance")) {
3218 debug_print_allowance = TRUE;
3219 } else if (!strcmp (opt, "print-pinning")) {
3220 sgen_pin_stats_enable ();
3221 } else if (!strcmp (opt, "verify-before-allocs")) {
3222 verify_before_allocs = 1;
3223 has_per_allocation_action = TRUE;
3224 } else if (g_str_has_prefix (opt, "verify-before-allocs=")) {
3225 char *arg = strchr (opt, '=') + 1;
3226 verify_before_allocs = atoi (arg);
3227 has_per_allocation_action = TRUE;
3228 } else if (!strcmp (opt, "collect-before-allocs")) {
3229 collect_before_allocs = 1;
3230 has_per_allocation_action = TRUE;
3231 } else if (g_str_has_prefix (opt, "collect-before-allocs=")) {
3232 char *arg = strchr (opt, '=') + 1;
3233 has_per_allocation_action = TRUE;
3234 collect_before_allocs = atoi (arg);
3235 } else if (!strcmp (opt, "verify-before-collections")) {
3236 whole_heap_check_before_collection = TRUE;
3237 } else if (!strcmp (opt, "check-remset-consistency")) {
3238 remset_consistency_checks = TRUE;
3239 nursery_clear_policy = CLEAR_AT_GC;
3240 } else if (!strcmp (opt, "mod-union-consistency-check")) {
3241 if (!major_collector.is_concurrent) {
3242 sgen_env_var_error (MONO_GC_DEBUG_NAME, "Ignoring.", "`mod-union-consistency-check` only works with concurrent major collector.");
3245 mod_union_consistency_check = TRUE;
3246 } else if (!strcmp (opt, "check-mark-bits")) {
3247 check_mark_bits_after_major_collection = TRUE;
3248 } else if (!strcmp (opt, "check-nursery-pinned")) {
3249 check_nursery_objects_pinned = TRUE;
3250 } else if (!strcmp (opt, "clear-at-gc")) {
3251 nursery_clear_policy = CLEAR_AT_GC;
3252 } else if (!strcmp (opt, "clear-nursery-at-gc")) {
3253 nursery_clear_policy = CLEAR_AT_GC;
3254 } else if (!strcmp (opt, "clear-at-tlab-creation")) {
3255 nursery_clear_policy = CLEAR_AT_TLAB_CREATION;
3256 } else if (!strcmp (opt, "debug-clear-at-tlab-creation")) {
3257 nursery_clear_policy = CLEAR_AT_TLAB_CREATION_DEBUG;
3258 } else if (!strcmp (opt, "check-scan-starts")) {
3259 do_scan_starts_check = TRUE;
3260 } else if (!strcmp (opt, "verify-nursery-at-minor-gc")) {
3261 do_verify_nursery = TRUE;
3262 } else if (!strcmp (opt, "check-concurrent")) {
3263 if (!major_collector.is_concurrent) {
3264 sgen_env_var_error (MONO_GC_DEBUG_NAME, "Ignoring.", "`check-concurrent` only works with concurrent major collectors.");
3267 nursery_clear_policy = CLEAR_AT_GC;
3268 do_concurrent_checks = TRUE;
3269 } else if (!strcmp (opt, "dump-nursery-at-minor-gc")) {
3270 do_dump_nursery_content = TRUE;
3271 } else if (!strcmp (opt, "disable-minor")) {
3272 disable_minor_collections = TRUE;
3273 } else if (!strcmp (opt, "disable-major")) {
3274 disable_major_collections = TRUE;
3275 } else if (g_str_has_prefix (opt, "heap-dump=")) {
3276 char *filename = strchr (opt, '=') + 1;
3277 nursery_clear_policy = CLEAR_AT_GC;
3278 sgen_debug_enable_heap_dump (filename);
3279 } else if (g_str_has_prefix (opt, "binary-protocol=")) {
3280 char *filename = strchr (opt, '=') + 1;
3281 char *colon = strrchr (filename, ':');
3284 if (!mono_gc_parse_environment_string_extract_number (colon + 1, &limit)) {
3285 sgen_env_var_error (MONO_GC_DEBUG_NAME, "Ignoring limit.", "Binary protocol file size limit must be an integer.");
3290 binary_protocol_init (filename, (long long)limit);
3291 } else if (!strcmp (opt, "nursery-canaries")) {
3292 do_verify_nursery = TRUE;
3293 enable_nursery_canaries = TRUE;
3294 } else if (!sgen_client_handle_gc_debug (opt)) {
3295 sgen_env_var_error (MONO_GC_DEBUG_NAME, "Ignoring.", "Unknown option `%s`.", opt);
3300 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);
3301 fprintf (stderr, "Valid <option>s are:\n");
3302 fprintf (stderr, " collect-before-allocs[=<n>]\n");
3303 fprintf (stderr, " verify-before-allocs[=<n>]\n");
3304 fprintf (stderr, " check-remset-consistency\n");
3305 fprintf (stderr, " check-mark-bits\n");
3306 fprintf (stderr, " check-nursery-pinned\n");
3307 fprintf (stderr, " verify-before-collections\n");
3308 fprintf (stderr, " verify-nursery-at-minor-gc\n");
3309 fprintf (stderr, " dump-nursery-at-minor-gc\n");
3310 fprintf (stderr, " disable-minor\n");
3311 fprintf (stderr, " disable-major\n");
3312 fprintf (stderr, " check-concurrent\n");
3313 fprintf (stderr, " clear-[nursery-]at-gc\n");
3314 fprintf (stderr, " clear-at-tlab-creation\n");
3315 fprintf (stderr, " debug-clear-at-tlab-creation\n");
3316 fprintf (stderr, " check-scan-starts\n");
3317 fprintf (stderr, " print-allowance\n");
3318 fprintf (stderr, " print-pinning\n");
3319 fprintf (stderr, " heap-dump=<filename>\n");
3320 fprintf (stderr, " binary-protocol=<filename>[:<file-size-limit>]\n");
3321 fprintf (stderr, " nursery-canaries\n");
3322 sgen_client_print_gc_debug_usage ();
3323 fprintf (stderr, "\n");
3325 usage_printed = TRUE;
3332 g_free (debug_opts);
3334 if (check_mark_bits_after_major_collection)
3335 nursery_clear_policy = CLEAR_AT_GC;
3337 if (major_collector.post_param_init)
3338 major_collector.post_param_init (&major_collector);
3340 if (major_collector.needs_thread_pool) {
3341 int num_workers = 1;
3342 if (major_collector.is_parallel) {
3343 /* FIXME Detect the number of physical cores, instead of logical */
3344 num_workers = mono_cpu_count () / 2;
3345 if (num_workers < 1)
3348 sgen_workers_init (num_workers, (SgenWorkerCallback) major_collector.worker_init_cb);
3351 sgen_memgov_init (max_heap, soft_limit, debug_print_allowance, allowance_ratio, save_target);
3353 memset (&remset, 0, sizeof (remset));
3355 sgen_card_table_init (&remset);
3357 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");
3361 sgen_init_bridge ();
3365 sgen_gc_initialized ()
3367 return gc_initialized > 0;
3371 sgen_get_nursery_clear_policy (void)
3373 return nursery_clear_policy;
3379 mono_coop_mutex_lock (&gc_mutex);
3383 sgen_gc_unlock (void)
3385 mono_coop_mutex_unlock (&gc_mutex);
3389 sgen_major_collector_iterate_live_block_ranges (sgen_cardtable_block_callback callback)
3391 major_collector.iterate_live_block_ranges (callback);
3395 sgen_major_collector_iterate_block_ranges (sgen_cardtable_block_callback callback)
3397 major_collector.iterate_block_ranges (callback);
3401 sgen_get_major_collector (void)
3403 return &major_collector;
3407 sgen_get_remset (void)
3413 count_cards (long long *major_total, long long *major_marked, long long *los_total, long long *los_marked)
3415 sgen_get_major_collector ()->count_cards (major_total, major_marked);
3416 sgen_los_count_cards (los_total, los_marked);
3419 static gboolean world_is_stopped = FALSE;
3421 /* LOCKING: assumes the GC lock is held */
3423 sgen_stop_world (int generation)
3425 long long major_total = -1, major_marked = -1, los_total = -1, los_marked = -1;
3427 SGEN_ASSERT (0, !world_is_stopped, "Why are we stopping a stopped world?");
3429 binary_protocol_world_stopping (generation, sgen_timestamp (), (gpointer) (gsize) mono_native_thread_id_get ());
3431 sgen_client_stop_world (generation);
3433 world_is_stopped = TRUE;
3435 if (binary_protocol_is_heavy_enabled ())
3436 count_cards (&major_total, &major_marked, &los_total, &los_marked);
3437 binary_protocol_world_stopped (generation, sgen_timestamp (), major_total, major_marked, los_total, los_marked);
3440 /* LOCKING: assumes the GC lock is held */
3442 sgen_restart_world (int generation)
3444 long long major_total = -1, major_marked = -1, los_total = -1, los_marked = -1;
3447 SGEN_ASSERT (0, world_is_stopped, "Why are we restarting a running world?");
3449 if (binary_protocol_is_heavy_enabled ())
3450 count_cards (&major_total, &major_marked, &los_total, &los_marked);
3451 binary_protocol_world_restarting (generation, sgen_timestamp (), major_total, major_marked, los_total, los_marked);
3453 world_is_stopped = FALSE;
3455 sgen_client_restart_world (generation, &stw_time);
3457 binary_protocol_world_restarted (generation, sgen_timestamp ());
3459 if (sgen_client_bridge_need_processing ())
3460 sgen_client_bridge_processing_finish (generation);
3462 sgen_memgov_collection_end (generation, stw_time);
3466 sgen_is_world_stopped (void)
3468 return world_is_stopped;
3472 sgen_check_whole_heap_stw (void)
3474 sgen_stop_world (0);
3475 sgen_clear_nursery_fragments ();
3476 sgen_check_whole_heap (TRUE);
3477 sgen_restart_world (0);
3481 sgen_timestamp (void)
3483 SGEN_TV_DECLARE (timestamp);
3484 SGEN_TV_GETTIME (timestamp);
3485 return SGEN_TV_ELAPSED (sgen_init_timestamp, timestamp);
3488 #endif /* HAVE_SGEN_GC */