1 /*
2 * kmp_alloc.cpp -- private/shared dynamic memory allocation and management
3 */
4
5 //===----------------------------------------------------------------------===//
6 //
7 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
8 // See https://llvm.org/LICENSE.txt for license information.
9 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
10 //
11 //===----------------------------------------------------------------------===//
12
13 #include "kmp.h"
14 #include "kmp_io.h"
15 #include "kmp_wrapper_malloc.h"
16
17 // Disable bget when it is not used
18 #if KMP_USE_BGET
19
20 /* Thread private buffer management code */
21
22 typedef int (*bget_compact_t)(size_t, int);
23 typedef void *(*bget_acquire_t)(size_t);
24 typedef void (*bget_release_t)(void *);
25
26 /* NOTE: bufsize must be a signed datatype */
27
28 #if KMP_OS_WINDOWS
29 #if KMP_ARCH_X86 || KMP_ARCH_ARM
30 typedef kmp_int32 bufsize;
31 #else
32 typedef kmp_int64 bufsize;
33 #endif
34 #else
35 typedef ssize_t bufsize;
36 #endif // KMP_OS_WINDOWS
37
38 /* The three modes of operation are, fifo search, lifo search, and best-fit */
39
40 typedef enum bget_mode {
41 bget_mode_fifo = 0,
42 bget_mode_lifo = 1,
43 bget_mode_best = 2
44 } bget_mode_t;
45
46 static void bpool(kmp_info_t *th, void *buffer, bufsize len);
47 static void *bget(kmp_info_t *th, bufsize size);
48 static void *bgetz(kmp_info_t *th, bufsize size);
49 static void *bgetr(kmp_info_t *th, void *buffer, bufsize newsize);
50 static void brel(kmp_info_t *th, void *buf);
51 static void bectl(kmp_info_t *th, bget_compact_t compact,
52 bget_acquire_t acquire, bget_release_t release,
53 bufsize pool_incr);
54
55 /* BGET CONFIGURATION */
56 /* Buffer allocation size quantum: all buffers allocated are a
57 multiple of this size. This MUST be a power of two. */
58
59 /* On IA-32 architecture with Linux* OS, malloc() does not
60 ensure 16 byte alignment */
61
62 #if KMP_ARCH_X86 || !KMP_HAVE_QUAD
63
64 #define SizeQuant 8
65 #define AlignType double
66
67 #else
68
69 #define SizeQuant 16
70 #define AlignType _Quad
71
72 #endif
73
74 // Define this symbol to enable the bstats() function which calculates the
75 // total free space in the buffer pool, the largest available buffer, and the
76 // total space currently allocated.
77 #define BufStats 1
78
79 #ifdef KMP_DEBUG
80
81 // Define this symbol to enable the bpoold() function which dumps the buffers
82 // in a buffer pool.
83 #define BufDump 1
84
85 // Define this symbol to enable the bpoolv() function for validating a buffer
86 // pool.
87 #define BufValid 1
88
89 // Define this symbol to enable the bufdump() function which allows dumping the
90 // contents of an allocated or free buffer.
91 #define DumpData 1
92
93 #ifdef NOT_USED_NOW
94
95 // Wipe free buffers to a guaranteed pattern of garbage to trip up miscreants
96 // who attempt to use pointers into released buffers.
97 #define FreeWipe 1
98
99 // Use a best fit algorithm when searching for space for an allocation request.
100 // This uses memory more efficiently, but allocation will be much slower.
101 #define BestFit 1
102
103 #endif /* NOT_USED_NOW */
104 #endif /* KMP_DEBUG */
105
106 static bufsize bget_bin_size[] = {
107 0,
108 // 1 << 6, /* .5 Cache line */
109 1 << 7, /* 1 Cache line, new */
110 1 << 8, /* 2 Cache lines */
111 1 << 9, /* 4 Cache lines, new */
112 1 << 10, /* 8 Cache lines */
113 1 << 11, /* 16 Cache lines, new */
114 1 << 12, 1 << 13, /* new */
115 1 << 14, 1 << 15, /* new */
116 1 << 16, 1 << 17, 1 << 18, 1 << 19, 1 << 20, /* 1MB */
117 1 << 21, /* 2MB */
118 1 << 22, /* 4MB */
119 1 << 23, /* 8MB */
120 1 << 24, /* 16MB */
121 1 << 25, /* 32MB */
122 };
123
124 #define MAX_BGET_BINS (int)(sizeof(bget_bin_size) / sizeof(bufsize))
125
126 struct bfhead;
127
128 // Declare the interface, including the requested buffer size type, bufsize.
129
130 /* Queue links */
131 typedef struct qlinks {
132 struct bfhead *flink; /* Forward link */
133 struct bfhead *blink; /* Backward link */
134 } qlinks_t;
135
136 /* Header in allocated and free buffers */
137 typedef struct bhead2 {
138 kmp_info_t *bthr; /* The thread which owns the buffer pool */
139 bufsize prevfree; /* Relative link back to previous free buffer in memory or
140 0 if previous buffer is allocated. */
141 bufsize bsize; /* Buffer size: positive if free, negative if allocated. */
142 } bhead2_t;
143
144 /* Make sure the bhead structure is a multiple of SizeQuant in size. */
145 typedef union bhead {
146 KMP_ALIGN(SizeQuant)
147 AlignType b_align;
148 char b_pad[sizeof(bhead2_t) + (SizeQuant - (sizeof(bhead2_t) % SizeQuant))];
149 bhead2_t bb;
150 } bhead_t;
151 #define BH(p) ((bhead_t *)(p))
152
153 /* Header in directly allocated buffers (by acqfcn) */
154 typedef struct bdhead {
155 bufsize tsize; /* Total size, including overhead */
156 bhead_t bh; /* Common header */
157 } bdhead_t;
158 #define BDH(p) ((bdhead_t *)(p))
159
160 /* Header in free buffers */
161 typedef struct bfhead {
162 bhead_t bh; /* Common allocated/free header */
163 qlinks_t ql; /* Links on free list */
164 } bfhead_t;
165 #define BFH(p) ((bfhead_t *)(p))
166
167 typedef struct thr_data {
168 bfhead_t freelist[MAX_BGET_BINS];
169 #if BufStats
170 size_t totalloc; /* Total space currently allocated */
171 long numget, numrel; /* Number of bget() and brel() calls */
172 long numpblk; /* Number of pool blocks */
173 long numpget, numprel; /* Number of block gets and rels */
174 long numdget, numdrel; /* Number of direct gets and rels */
175 #endif /* BufStats */
176
177 /* Automatic expansion block management functions */
178 bget_compact_t compfcn;
179 bget_acquire_t acqfcn;
180 bget_release_t relfcn;
181
182 bget_mode_t mode; /* what allocation mode to use? */
183
184 bufsize exp_incr; /* Expansion block size */
185 bufsize pool_len; /* 0: no bpool calls have been made
186 -1: not all pool blocks are the same size
187 >0: (common) block size for all bpool calls made so far
188 */
189 bfhead_t *last_pool; /* Last pool owned by this thread (delay deallocation) */
190 } thr_data_t;
191
192 /* Minimum allocation quantum: */
193 #define QLSize (sizeof(qlinks_t))
194 #define SizeQ ((SizeQuant > QLSize) ? SizeQuant : QLSize)
195 #define MaxSize \
196 (bufsize)( \
197 ~(((bufsize)(1) << (sizeof(bufsize) * CHAR_BIT - 1)) | (SizeQuant - 1)))
198 // Maximum for the requested size.
199
200 /* End sentinel: value placed in bsize field of dummy block delimiting
201 end of pool block. The most negative number which will fit in a
202 bufsize, defined in a way that the compiler will accept. */
203
204 #define ESent \
205 ((bufsize)(-(((((bufsize)1) << ((int)sizeof(bufsize) * 8 - 2)) - 1) * 2) - 2))
206
207 /* Thread Data management routines */
bget_get_bin(bufsize size)208 static int bget_get_bin(bufsize size) {
209 // binary chop bins
210 int lo = 0, hi = MAX_BGET_BINS - 1;
211
212 KMP_DEBUG_ASSERT(size > 0);
213
214 while ((hi - lo) > 1) {
215 int mid = (lo + hi) >> 1;
216 if (size < bget_bin_size[mid])
217 hi = mid - 1;
218 else
219 lo = mid;
220 }
221
222 KMP_DEBUG_ASSERT((lo >= 0) && (lo < MAX_BGET_BINS));
223
224 return lo;
225 }
226
set_thr_data(kmp_info_t * th)227 static void set_thr_data(kmp_info_t *th) {
228 int i;
229 thr_data_t *data;
230
231 data = (thr_data_t *)((!th->th.th_local.bget_data)
232 ? __kmp_allocate(sizeof(*data))
233 : th->th.th_local.bget_data);
234
235 memset(data, '\0', sizeof(*data));
236
237 for (i = 0; i < MAX_BGET_BINS; ++i) {
238 data->freelist[i].ql.flink = &data->freelist[i];
239 data->freelist[i].ql.blink = &data->freelist[i];
240 }
241
242 th->th.th_local.bget_data = data;
243 th->th.th_local.bget_list = 0;
244 #if !USE_CMP_XCHG_FOR_BGET
245 #ifdef USE_QUEUING_LOCK_FOR_BGET
246 __kmp_init_lock(&th->th.th_local.bget_lock);
247 #else
248 __kmp_init_bootstrap_lock(&th->th.th_local.bget_lock);
249 #endif /* USE_LOCK_FOR_BGET */
250 #endif /* ! USE_CMP_XCHG_FOR_BGET */
251 }
252
get_thr_data(kmp_info_t * th)253 static thr_data_t *get_thr_data(kmp_info_t *th) {
254 thr_data_t *data;
255
256 data = (thr_data_t *)th->th.th_local.bget_data;
257
258 KMP_DEBUG_ASSERT(data != 0);
259
260 return data;
261 }
262
263 /* Walk the free list and release the enqueued buffers */
__kmp_bget_dequeue(kmp_info_t * th)264 static void __kmp_bget_dequeue(kmp_info_t *th) {
265 void *p = TCR_SYNC_PTR(th->th.th_local.bget_list);
266
267 if (p != 0) {
268 #if USE_CMP_XCHG_FOR_BGET
269 {
270 volatile void *old_value = TCR_SYNC_PTR(th->th.th_local.bget_list);
271 while (!KMP_COMPARE_AND_STORE_PTR(&th->th.th_local.bget_list,
272 CCAST(void *, old_value), nullptr)) {
273 KMP_CPU_PAUSE();
274 old_value = TCR_SYNC_PTR(th->th.th_local.bget_list);
275 }
276 p = CCAST(void *, old_value);
277 }
278 #else /* ! USE_CMP_XCHG_FOR_BGET */
279 #ifdef USE_QUEUING_LOCK_FOR_BGET
280 __kmp_acquire_lock(&th->th.th_local.bget_lock, __kmp_gtid_from_thread(th));
281 #else
282 __kmp_acquire_bootstrap_lock(&th->th.th_local.bget_lock);
283 #endif /* USE_QUEUING_LOCK_FOR_BGET */
284
285 p = (void *)th->th.th_local.bget_list;
286 th->th.th_local.bget_list = 0;
287
288 #ifdef USE_QUEUING_LOCK_FOR_BGET
289 __kmp_release_lock(&th->th.th_local.bget_lock, __kmp_gtid_from_thread(th));
290 #else
291 __kmp_release_bootstrap_lock(&th->th.th_local.bget_lock);
292 #endif
293 #endif /* USE_CMP_XCHG_FOR_BGET */
294
295 /* Check again to make sure the list is not empty */
296 while (p != 0) {
297 void *buf = p;
298 bfhead_t *b = BFH(((char *)p) - sizeof(bhead_t));
299
300 KMP_DEBUG_ASSERT(b->bh.bb.bsize != 0);
301 KMP_DEBUG_ASSERT(((kmp_uintptr_t)TCR_PTR(b->bh.bb.bthr) & ~1) ==
302 (kmp_uintptr_t)th); // clear possible mark
303 KMP_DEBUG_ASSERT(b->ql.blink == 0);
304
305 p = (void *)b->ql.flink;
306
307 brel(th, buf);
308 }
309 }
310 }
311
312 /* Chain together the free buffers by using the thread owner field */
__kmp_bget_enqueue(kmp_info_t * th,void * buf,kmp_int32 rel_gtid)313 static void __kmp_bget_enqueue(kmp_info_t *th, void *buf
314 #ifdef USE_QUEUING_LOCK_FOR_BGET
315 ,
316 kmp_int32 rel_gtid
317 #endif
318 ) {
319 bfhead_t *b = BFH(((char *)buf) - sizeof(bhead_t));
320
321 KMP_DEBUG_ASSERT(b->bh.bb.bsize != 0);
322 KMP_DEBUG_ASSERT(((kmp_uintptr_t)TCR_PTR(b->bh.bb.bthr) & ~1) ==
323 (kmp_uintptr_t)th); // clear possible mark
324
325 b->ql.blink = 0;
326
327 KC_TRACE(10, ("__kmp_bget_enqueue: moving buffer to T#%d list\n",
328 __kmp_gtid_from_thread(th)));
329
330 #if USE_CMP_XCHG_FOR_BGET
331 {
332 volatile void *old_value = TCR_PTR(th->th.th_local.bget_list);
333 /* the next pointer must be set before setting bget_list to buf to avoid
334 exposing a broken list to other threads, even for an instant. */
335 b->ql.flink = BFH(CCAST(void *, old_value));
336
337 while (!KMP_COMPARE_AND_STORE_PTR(&th->th.th_local.bget_list,
338 CCAST(void *, old_value), buf)) {
339 KMP_CPU_PAUSE();
340 old_value = TCR_PTR(th->th.th_local.bget_list);
341 /* the next pointer must be set before setting bget_list to buf to avoid
342 exposing a broken list to other threads, even for an instant. */
343 b->ql.flink = BFH(CCAST(void *, old_value));
344 }
345 }
346 #else /* ! USE_CMP_XCHG_FOR_BGET */
347 #ifdef USE_QUEUING_LOCK_FOR_BGET
348 __kmp_acquire_lock(&th->th.th_local.bget_lock, rel_gtid);
349 #else
350 __kmp_acquire_bootstrap_lock(&th->th.th_local.bget_lock);
351 #endif
352
353 b->ql.flink = BFH(th->th.th_local.bget_list);
354 th->th.th_local.bget_list = (void *)buf;
355
356 #ifdef USE_QUEUING_LOCK_FOR_BGET
357 __kmp_release_lock(&th->th.th_local.bget_lock, rel_gtid);
358 #else
359 __kmp_release_bootstrap_lock(&th->th.th_local.bget_lock);
360 #endif
361 #endif /* USE_CMP_XCHG_FOR_BGET */
362 }
363
364 /* insert buffer back onto a new freelist */
__kmp_bget_insert_into_freelist(thr_data_t * thr,bfhead_t * b)365 static void __kmp_bget_insert_into_freelist(thr_data_t *thr, bfhead_t *b) {
366 int bin;
367
368 KMP_DEBUG_ASSERT(((size_t)b) % SizeQuant == 0);
369 KMP_DEBUG_ASSERT(b->bh.bb.bsize % SizeQuant == 0);
370
371 bin = bget_get_bin(b->bh.bb.bsize);
372
373 KMP_DEBUG_ASSERT(thr->freelist[bin].ql.blink->ql.flink ==
374 &thr->freelist[bin]);
375 KMP_DEBUG_ASSERT(thr->freelist[bin].ql.flink->ql.blink ==
376 &thr->freelist[bin]);
377
378 b->ql.flink = &thr->freelist[bin];
379 b->ql.blink = thr->freelist[bin].ql.blink;
380
381 thr->freelist[bin].ql.blink = b;
382 b->ql.blink->ql.flink = b;
383 }
384
385 /* unlink the buffer from the old freelist */
__kmp_bget_remove_from_freelist(bfhead_t * b)386 static void __kmp_bget_remove_from_freelist(bfhead_t *b) {
387 KMP_DEBUG_ASSERT(b->ql.blink->ql.flink == b);
388 KMP_DEBUG_ASSERT(b->ql.flink->ql.blink == b);
389
390 b->ql.blink->ql.flink = b->ql.flink;
391 b->ql.flink->ql.blink = b->ql.blink;
392 }
393
394 /* GET STATS -- check info on free list */
bcheck(kmp_info_t * th,bufsize * max_free,bufsize * total_free)395 static void bcheck(kmp_info_t *th, bufsize *max_free, bufsize *total_free) {
396 thr_data_t *thr = get_thr_data(th);
397 int bin;
398
399 *total_free = *max_free = 0;
400
401 for (bin = 0; bin < MAX_BGET_BINS; ++bin) {
402 bfhead_t *b, *best;
403
404 best = &thr->freelist[bin];
405 b = best->ql.flink;
406
407 while (b != &thr->freelist[bin]) {
408 *total_free += (b->bh.bb.bsize - sizeof(bhead_t));
409 if ((best == &thr->freelist[bin]) || (b->bh.bb.bsize < best->bh.bb.bsize))
410 best = b;
411
412 /* Link to next buffer */
413 b = b->ql.flink;
414 }
415
416 if (*max_free < best->bh.bb.bsize)
417 *max_free = best->bh.bb.bsize;
418 }
419
420 if (*max_free > (bufsize)sizeof(bhead_t))
421 *max_free -= sizeof(bhead_t);
422 }
423
424 /* BGET -- Allocate a buffer. */
bget(kmp_info_t * th,bufsize requested_size)425 static void *bget(kmp_info_t *th, bufsize requested_size) {
426 thr_data_t *thr = get_thr_data(th);
427 bufsize size = requested_size;
428 bfhead_t *b;
429 void *buf;
430 int compactseq = 0;
431 int use_blink = 0;
432 /* For BestFit */
433 bfhead_t *best;
434
435 if (size < 0 || size + sizeof(bhead_t) > MaxSize) {
436 return NULL;
437 }
438
439 __kmp_bget_dequeue(th); /* Release any queued buffers */
440
441 if (size < (bufsize)SizeQ) { // Need at least room for the queue links.
442 size = SizeQ;
443 }
444 #if defined(SizeQuant) && (SizeQuant > 1)
445 size = (size + (SizeQuant - 1)) & (~(SizeQuant - 1));
446 #endif
447
448 size += sizeof(bhead_t); // Add overhead in allocated buffer to size required.
449 KMP_DEBUG_ASSERT(size >= 0);
450 KMP_DEBUG_ASSERT(size % SizeQuant == 0);
451
452 use_blink = (thr->mode == bget_mode_lifo);
453
454 /* If a compact function was provided in the call to bectl(), wrap
455 a loop around the allocation process to allow compaction to
456 intervene in case we don't find a suitable buffer in the chain. */
457
458 for (;;) {
459 int bin;
460
461 for (bin = bget_get_bin(size); bin < MAX_BGET_BINS; ++bin) {
462 /* Link to next buffer */
463 b = (use_blink ? thr->freelist[bin].ql.blink
464 : thr->freelist[bin].ql.flink);
465
466 if (thr->mode == bget_mode_best) {
467 best = &thr->freelist[bin];
468
469 /* Scan the free list searching for the first buffer big enough
470 to hold the requested size buffer. */
471 while (b != &thr->freelist[bin]) {
472 if (b->bh.bb.bsize >= (bufsize)size) {
473 if ((best == &thr->freelist[bin]) ||
474 (b->bh.bb.bsize < best->bh.bb.bsize)) {
475 best = b;
476 }
477 }
478
479 /* Link to next buffer */
480 b = (use_blink ? b->ql.blink : b->ql.flink);
481 }
482 b = best;
483 }
484
485 while (b != &thr->freelist[bin]) {
486 if ((bufsize)b->bh.bb.bsize >= (bufsize)size) {
487
488 // Buffer is big enough to satisfy the request. Allocate it to the
489 // caller. We must decide whether the buffer is large enough to split
490 // into the part given to the caller and a free buffer that remains
491 // on the free list, or whether the entire buffer should be removed
492 // from the free list and given to the caller in its entirety. We
493 // only split the buffer if enough room remains for a header plus the
494 // minimum quantum of allocation.
495 if ((b->bh.bb.bsize - (bufsize)size) >
496 (bufsize)(SizeQ + (sizeof(bhead_t)))) {
497 bhead_t *ba, *bn;
498
499 ba = BH(((char *)b) + (b->bh.bb.bsize - (bufsize)size));
500 bn = BH(((char *)ba) + size);
501
502 KMP_DEBUG_ASSERT(bn->bb.prevfree == b->bh.bb.bsize);
503
504 /* Subtract size from length of free block. */
505 b->bh.bb.bsize -= (bufsize)size;
506
507 /* Link allocated buffer to the previous free buffer. */
508 ba->bb.prevfree = b->bh.bb.bsize;
509
510 /* Plug negative size into user buffer. */
511 ba->bb.bsize = -size;
512
513 /* Mark this buffer as owned by this thread. */
514 TCW_PTR(ba->bb.bthr,
515 th); // not an allocated address (do not mark it)
516 /* Mark buffer after this one not preceded by free block. */
517 bn->bb.prevfree = 0;
518
519 // unlink buffer from old freelist, and reinsert into new freelist
520 __kmp_bget_remove_from_freelist(b);
521 __kmp_bget_insert_into_freelist(thr, b);
522 #if BufStats
523 thr->totalloc += (size_t)size;
524 thr->numget++; /* Increment number of bget() calls */
525 #endif
526 buf = (void *)((((char *)ba) + sizeof(bhead_t)));
527 KMP_DEBUG_ASSERT(((size_t)buf) % SizeQuant == 0);
528 return buf;
529 } else {
530 bhead_t *ba;
531
532 ba = BH(((char *)b) + b->bh.bb.bsize);
533
534 KMP_DEBUG_ASSERT(ba->bb.prevfree == b->bh.bb.bsize);
535
536 /* The buffer isn't big enough to split. Give the whole
537 shebang to the caller and remove it from the free list. */
538
539 __kmp_bget_remove_from_freelist(b);
540 #if BufStats
541 thr->totalloc += (size_t)b->bh.bb.bsize;
542 thr->numget++; /* Increment number of bget() calls */
543 #endif
544 /* Negate size to mark buffer allocated. */
545 b->bh.bb.bsize = -(b->bh.bb.bsize);
546
547 /* Mark this buffer as owned by this thread. */
548 TCW_PTR(ba->bb.bthr, th); // not an allocated address (do not mark)
549 /* Zero the back pointer in the next buffer in memory
550 to indicate that this buffer is allocated. */
551 ba->bb.prevfree = 0;
552
553 /* Give user buffer starting at queue links. */
554 buf = (void *)&(b->ql);
555 KMP_DEBUG_ASSERT(((size_t)buf) % SizeQuant == 0);
556 return buf;
557 }
558 }
559
560 /* Link to next buffer */
561 b = (use_blink ? b->ql.blink : b->ql.flink);
562 }
563 }
564
565 /* We failed to find a buffer. If there's a compact function defined,
566 notify it of the size requested. If it returns TRUE, try the allocation
567 again. */
568
569 if ((thr->compfcn == 0) || (!(*thr->compfcn)(size, ++compactseq))) {
570 break;
571 }
572 }
573
574 /* No buffer available with requested size free. */
575
576 /* Don't give up yet -- look in the reserve supply. */
577 if (thr->acqfcn != 0) {
578 if (size > (bufsize)(thr->exp_incr - sizeof(bhead_t))) {
579 /* Request is too large to fit in a single expansion block.
580 Try to satisfy it by a direct buffer acquisition. */
581 bdhead_t *bdh;
582
583 size += sizeof(bdhead_t) - sizeof(bhead_t);
584
585 KE_TRACE(10, ("%%%%%% MALLOC( %d )\n", (int)size));
586
587 /* richryan */
588 bdh = BDH((*thr->acqfcn)((bufsize)size));
589 if (bdh != NULL) {
590
591 // Mark the buffer special by setting size field of its header to zero.
592 bdh->bh.bb.bsize = 0;
593
594 /* Mark this buffer as owned by this thread. */
595 TCW_PTR(bdh->bh.bb.bthr, th); // don't mark buffer as allocated,
596 // because direct buffer never goes to free list
597 bdh->bh.bb.prevfree = 0;
598 bdh->tsize = size;
599 #if BufStats
600 thr->totalloc += (size_t)size;
601 thr->numget++; /* Increment number of bget() calls */
602 thr->numdget++; /* Direct bget() call count */
603 #endif
604 buf = (void *)(bdh + 1);
605 KMP_DEBUG_ASSERT(((size_t)buf) % SizeQuant == 0);
606 return buf;
607 }
608
609 } else {
610
611 /* Try to obtain a new expansion block */
612 void *newpool;
613
614 KE_TRACE(10, ("%%%%%% MALLOCB( %d )\n", (int)thr->exp_incr));
615
616 /* richryan */
617 newpool = (*thr->acqfcn)((bufsize)thr->exp_incr);
618 KMP_DEBUG_ASSERT(((size_t)newpool) % SizeQuant == 0);
619 if (newpool != NULL) {
620 bpool(th, newpool, thr->exp_incr);
621 buf = bget(
622 th, requested_size); /* This can't, I say, can't get into a loop. */
623 return buf;
624 }
625 }
626 }
627
628 /* Still no buffer available */
629
630 return NULL;
631 }
632
633 /* BGETZ -- Allocate a buffer and clear its contents to zero. We clear
634 the entire contents of the buffer to zero, not just the
635 region requested by the caller. */
636
bgetz(kmp_info_t * th,bufsize size)637 static void *bgetz(kmp_info_t *th, bufsize size) {
638 char *buf = (char *)bget(th, size);
639
640 if (buf != NULL) {
641 bhead_t *b;
642 bufsize rsize;
643
644 b = BH(buf - sizeof(bhead_t));
645 rsize = -(b->bb.bsize);
646 if (rsize == 0) {
647 bdhead_t *bd;
648
649 bd = BDH(buf - sizeof(bdhead_t));
650 rsize = bd->tsize - (bufsize)sizeof(bdhead_t);
651 } else {
652 rsize -= sizeof(bhead_t);
653 }
654
655 KMP_DEBUG_ASSERT(rsize >= size);
656
657 (void)memset(buf, 0, (bufsize)rsize);
658 }
659 return ((void *)buf);
660 }
661
662 /* BGETR -- Reallocate a buffer. This is a minimal implementation,
663 simply in terms of brel() and bget(). It could be
664 enhanced to allow the buffer to grow into adjacent free
665 blocks and to avoid moving data unnecessarily. */
666
bgetr(kmp_info_t * th,void * buf,bufsize size)667 static void *bgetr(kmp_info_t *th, void *buf, bufsize size) {
668 void *nbuf;
669 bufsize osize; /* Old size of buffer */
670 bhead_t *b;
671
672 nbuf = bget(th, size);
673 if (nbuf == NULL) { /* Acquire new buffer */
674 return NULL;
675 }
676 if (buf == NULL) {
677 return nbuf;
678 }
679 b = BH(((char *)buf) - sizeof(bhead_t));
680 osize = -b->bb.bsize;
681 if (osize == 0) {
682 /* Buffer acquired directly through acqfcn. */
683 bdhead_t *bd;
684
685 bd = BDH(((char *)buf) - sizeof(bdhead_t));
686 osize = bd->tsize - (bufsize)sizeof(bdhead_t);
687 } else {
688 osize -= sizeof(bhead_t);
689 }
690
691 KMP_DEBUG_ASSERT(osize > 0);
692
693 (void)KMP_MEMCPY((char *)nbuf, (char *)buf, /* Copy the data */
694 (size_t)((size < osize) ? size : osize));
695 brel(th, buf);
696
697 return nbuf;
698 }
699
700 /* BREL -- Release a buffer. */
brel(kmp_info_t * th,void * buf)701 static void brel(kmp_info_t *th, void *buf) {
702 thr_data_t *thr = get_thr_data(th);
703 bfhead_t *b, *bn;
704 kmp_info_t *bth;
705
706 KMP_DEBUG_ASSERT(buf != NULL);
707 KMP_DEBUG_ASSERT(((size_t)buf) % SizeQuant == 0);
708
709 b = BFH(((char *)buf) - sizeof(bhead_t));
710
711 if (b->bh.bb.bsize == 0) { /* Directly-acquired buffer? */
712 bdhead_t *bdh;
713
714 bdh = BDH(((char *)buf) - sizeof(bdhead_t));
715 KMP_DEBUG_ASSERT(b->bh.bb.prevfree == 0);
716 #if BufStats
717 thr->totalloc -= (size_t)bdh->tsize;
718 thr->numdrel++; /* Number of direct releases */
719 thr->numrel++; /* Increment number of brel() calls */
720 #endif /* BufStats */
721 #ifdef FreeWipe
722 (void)memset((char *)buf, 0x55, (size_t)(bdh->tsize - sizeof(bdhead_t)));
723 #endif /* FreeWipe */
724
725 KE_TRACE(10, ("%%%%%% FREE( %p )\n", (void *)bdh));
726
727 KMP_DEBUG_ASSERT(thr->relfcn != 0);
728 (*thr->relfcn)((void *)bdh); /* Release it directly. */
729 return;
730 }
731
732 bth = (kmp_info_t *)((kmp_uintptr_t)TCR_PTR(b->bh.bb.bthr) &
733 ~1); // clear possible mark before comparison
734 if (bth != th) {
735 /* Add this buffer to be released by the owning thread later */
736 __kmp_bget_enqueue(bth, buf
737 #ifdef USE_QUEUING_LOCK_FOR_BGET
738 ,
739 __kmp_gtid_from_thread(th)
740 #endif
741 );
742 return;
743 }
744
745 /* Buffer size must be negative, indicating that the buffer is allocated. */
746 if (b->bh.bb.bsize >= 0) {
747 bn = NULL;
748 }
749 KMP_DEBUG_ASSERT(b->bh.bb.bsize < 0);
750
751 /* Back pointer in next buffer must be zero, indicating the same thing: */
752
753 KMP_DEBUG_ASSERT(BH((char *)b - b->bh.bb.bsize)->bb.prevfree == 0);
754
755 #if BufStats
756 thr->numrel++; /* Increment number of brel() calls */
757 thr->totalloc += (size_t)b->bh.bb.bsize;
758 #endif
759
760 /* If the back link is nonzero, the previous buffer is free. */
761
762 if (b->bh.bb.prevfree != 0) {
763 /* The previous buffer is free. Consolidate this buffer with it by adding
764 the length of this buffer to the previous free buffer. Note that we
765 subtract the size in the buffer being released, since it's negative to
766 indicate that the buffer is allocated. */
767 bufsize size = b->bh.bb.bsize;
768
769 /* Make the previous buffer the one we're working on. */
770 KMP_DEBUG_ASSERT(BH((char *)b - b->bh.bb.prevfree)->bb.bsize ==
771 b->bh.bb.prevfree);
772 b = BFH(((char *)b) - b->bh.bb.prevfree);
773 b->bh.bb.bsize -= size;
774
775 /* unlink the buffer from the old freelist */
776 __kmp_bget_remove_from_freelist(b);
777 } else {
778 /* The previous buffer isn't allocated. Mark this buffer size as positive
779 (i.e. free) and fall through to place the buffer on the free list as an
780 isolated free block. */
781 b->bh.bb.bsize = -b->bh.bb.bsize;
782 }
783
784 /* insert buffer back onto a new freelist */
785 __kmp_bget_insert_into_freelist(thr, b);
786
787 /* Now we look at the next buffer in memory, located by advancing from
788 the start of this buffer by its size, to see if that buffer is
789 free. If it is, we combine this buffer with the next one in
790 memory, dechaining the second buffer from the free list. */
791 bn = BFH(((char *)b) + b->bh.bb.bsize);
792 if (bn->bh.bb.bsize > 0) {
793
794 /* The buffer is free. Remove it from the free list and add
795 its size to that of our buffer. */
796 KMP_DEBUG_ASSERT(BH((char *)bn + bn->bh.bb.bsize)->bb.prevfree ==
797 bn->bh.bb.bsize);
798
799 __kmp_bget_remove_from_freelist(bn);
800
801 b->bh.bb.bsize += bn->bh.bb.bsize;
802
803 /* unlink the buffer from the old freelist, and reinsert it into the new
804 * freelist */
805 __kmp_bget_remove_from_freelist(b);
806 __kmp_bget_insert_into_freelist(thr, b);
807
808 /* Finally, advance to the buffer that follows the newly
809 consolidated free block. We must set its backpointer to the
810 head of the consolidated free block. We know the next block
811 must be an allocated block because the process of recombination
812 guarantees that two free blocks will never be contiguous in
813 memory. */
814 bn = BFH(((char *)b) + b->bh.bb.bsize);
815 }
816 #ifdef FreeWipe
817 (void)memset(((char *)b) + sizeof(bfhead_t), 0x55,
818 (size_t)(b->bh.bb.bsize - sizeof(bfhead_t)));
819 #endif
820 KMP_DEBUG_ASSERT(bn->bh.bb.bsize < 0);
821
822 /* The next buffer is allocated. Set the backpointer in it to point
823 to this buffer; the previous free buffer in memory. */
824
825 bn->bh.bb.prevfree = b->bh.bb.bsize;
826
827 /* If a block-release function is defined, and this free buffer
828 constitutes the entire block, release it. Note that pool_len
829 is defined in such a way that the test will fail unless all
830 pool blocks are the same size. */
831 if (thr->relfcn != 0 &&
832 b->bh.bb.bsize == (bufsize)(thr->pool_len - sizeof(bhead_t))) {
833 #if BufStats
834 if (thr->numpblk !=
835 1) { /* Do not release the last buffer until finalization time */
836 #endif
837
838 KMP_DEBUG_ASSERT(b->bh.bb.prevfree == 0);
839 KMP_DEBUG_ASSERT(BH((char *)b + b->bh.bb.bsize)->bb.bsize == ESent);
840 KMP_DEBUG_ASSERT(BH((char *)b + b->bh.bb.bsize)->bb.prevfree ==
841 b->bh.bb.bsize);
842
843 /* Unlink the buffer from the free list */
844 __kmp_bget_remove_from_freelist(b);
845
846 KE_TRACE(10, ("%%%%%% FREE( %p )\n", (void *)b));
847
848 (*thr->relfcn)(b);
849 #if BufStats
850 thr->numprel++; /* Nr of expansion block releases */
851 thr->numpblk--; /* Total number of blocks */
852 KMP_DEBUG_ASSERT(thr->numpblk == thr->numpget - thr->numprel);
853
854 // avoid leaving stale last_pool pointer around if it is being dealloced
855 if (thr->last_pool == b)
856 thr->last_pool = 0;
857 } else {
858 thr->last_pool = b;
859 }
860 #endif /* BufStats */
861 }
862 }
863
864 /* BECTL -- Establish automatic pool expansion control */
bectl(kmp_info_t * th,bget_compact_t compact,bget_acquire_t acquire,bget_release_t release,bufsize pool_incr)865 static void bectl(kmp_info_t *th, bget_compact_t compact,
866 bget_acquire_t acquire, bget_release_t release,
867 bufsize pool_incr) {
868 thr_data_t *thr = get_thr_data(th);
869
870 thr->compfcn = compact;
871 thr->acqfcn = acquire;
872 thr->relfcn = release;
873 thr->exp_incr = pool_incr;
874 }
875
876 /* BPOOL -- Add a region of memory to the buffer pool. */
bpool(kmp_info_t * th,void * buf,bufsize len)877 static void bpool(kmp_info_t *th, void *buf, bufsize len) {
878 /* int bin = 0; */
879 thr_data_t *thr = get_thr_data(th);
880 bfhead_t *b = BFH(buf);
881 bhead_t *bn;
882
883 __kmp_bget_dequeue(th); /* Release any queued buffers */
884
885 #ifdef SizeQuant
886 len &= ~(SizeQuant - 1);
887 #endif
888 if (thr->pool_len == 0) {
889 thr->pool_len = len;
890 } else if (len != thr->pool_len) {
891 thr->pool_len = -1;
892 }
893 #if BufStats
894 thr->numpget++; /* Number of block acquisitions */
895 thr->numpblk++; /* Number of blocks total */
896 KMP_DEBUG_ASSERT(thr->numpblk == thr->numpget - thr->numprel);
897 #endif /* BufStats */
898
899 /* Since the block is initially occupied by a single free buffer,
900 it had better not be (much) larger than the largest buffer
901 whose size we can store in bhead.bb.bsize. */
902 KMP_DEBUG_ASSERT(len - sizeof(bhead_t) <= -((bufsize)ESent + 1));
903
904 /* Clear the backpointer at the start of the block to indicate that
905 there is no free block prior to this one. That blocks
906 recombination when the first block in memory is released. */
907 b->bh.bb.prevfree = 0;
908
909 /* Create a dummy allocated buffer at the end of the pool. This dummy
910 buffer is seen when a buffer at the end of the pool is released and
911 blocks recombination of the last buffer with the dummy buffer at
912 the end. The length in the dummy buffer is set to the largest
913 negative number to denote the end of the pool for diagnostic
914 routines (this specific value is not counted on by the actual
915 allocation and release functions). */
916 len -= sizeof(bhead_t);
917 b->bh.bb.bsize = (bufsize)len;
918 /* Set the owner of this buffer */
919 TCW_PTR(b->bh.bb.bthr,
920 (kmp_info_t *)((kmp_uintptr_t)th |
921 1)); // mark the buffer as allocated address
922
923 /* Chain the new block to the free list. */
924 __kmp_bget_insert_into_freelist(thr, b);
925
926 #ifdef FreeWipe
927 (void)memset(((char *)b) + sizeof(bfhead_t), 0x55,
928 (size_t)(len - sizeof(bfhead_t)));
929 #endif
930 bn = BH(((char *)b) + len);
931 bn->bb.prevfree = (bufsize)len;
932 /* Definition of ESent assumes two's complement! */
933 KMP_DEBUG_ASSERT((~0) == -1 && (bn != 0));
934
935 bn->bb.bsize = ESent;
936 }
937
938 /* BFREED -- Dump the free lists for this thread. */
bfreed(kmp_info_t * th)939 static void bfreed(kmp_info_t *th) {
940 int bin = 0, count = 0;
941 int gtid = __kmp_gtid_from_thread(th);
942 thr_data_t *thr = get_thr_data(th);
943
944 #if BufStats
945 __kmp_printf_no_lock("__kmp_printpool: T#%d total=%" KMP_UINT64_SPEC
946 " get=%" KMP_INT64_SPEC " rel=%" KMP_INT64_SPEC
947 " pblk=%" KMP_INT64_SPEC " pget=%" KMP_INT64_SPEC
948 " prel=%" KMP_INT64_SPEC " dget=%" KMP_INT64_SPEC
949 " drel=%" KMP_INT64_SPEC "\n",
950 gtid, (kmp_uint64)thr->totalloc, (kmp_int64)thr->numget,
951 (kmp_int64)thr->numrel, (kmp_int64)thr->numpblk,
952 (kmp_int64)thr->numpget, (kmp_int64)thr->numprel,
953 (kmp_int64)thr->numdget, (kmp_int64)thr->numdrel);
954 #endif
955
956 for (bin = 0; bin < MAX_BGET_BINS; ++bin) {
957 bfhead_t *b;
958
959 for (b = thr->freelist[bin].ql.flink; b != &thr->freelist[bin];
960 b = b->ql.flink) {
961 bufsize bs = b->bh.bb.bsize;
962
963 KMP_DEBUG_ASSERT(b->ql.blink->ql.flink == b);
964 KMP_DEBUG_ASSERT(b->ql.flink->ql.blink == b);
965 KMP_DEBUG_ASSERT(bs > 0);
966
967 count += 1;
968
969 __kmp_printf_no_lock(
970 "__kmp_printpool: T#%d Free block: 0x%p size %6ld bytes.\n", gtid, b,
971 (long)bs);
972 #ifdef FreeWipe
973 {
974 char *lerr = ((char *)b) + sizeof(bfhead_t);
975 if ((bs > sizeof(bfhead_t)) &&
976 ((*lerr != 0x55) ||
977 (memcmp(lerr, lerr + 1, (size_t)(bs - (sizeof(bfhead_t) + 1))) !=
978 0))) {
979 __kmp_printf_no_lock("__kmp_printpool: T#%d (Contents of above "
980 "free block have been overstored.)\n",
981 gtid);
982 }
983 }
984 #endif
985 }
986 }
987
988 if (count == 0)
989 __kmp_printf_no_lock("__kmp_printpool: T#%d No free blocks\n", gtid);
990 }
991
__kmp_initialize_bget(kmp_info_t * th)992 void __kmp_initialize_bget(kmp_info_t *th) {
993 KMP_DEBUG_ASSERT(SizeQuant >= sizeof(void *) && (th != 0));
994
995 set_thr_data(th);
996
997 bectl(th, (bget_compact_t)0, (bget_acquire_t)malloc, (bget_release_t)free,
998 (bufsize)__kmp_malloc_pool_incr);
999 }
1000
__kmp_finalize_bget(kmp_info_t * th)1001 void __kmp_finalize_bget(kmp_info_t *th) {
1002 thr_data_t *thr;
1003 bfhead_t *b;
1004
1005 KMP_DEBUG_ASSERT(th != 0);
1006
1007 #if BufStats
1008 thr = (thr_data_t *)th->th.th_local.bget_data;
1009 KMP_DEBUG_ASSERT(thr != NULL);
1010 b = thr->last_pool;
1011
1012 /* If a block-release function is defined, and this free buffer constitutes
1013 the entire block, release it. Note that pool_len is defined in such a way
1014 that the test will fail unless all pool blocks are the same size. */
1015
1016 // Deallocate the last pool if one exists because we no longer do it in brel()
1017 if (thr->relfcn != 0 && b != 0 && thr->numpblk != 0 &&
1018 b->bh.bb.bsize == (bufsize)(thr->pool_len - sizeof(bhead_t))) {
1019 KMP_DEBUG_ASSERT(b->bh.bb.prevfree == 0);
1020 KMP_DEBUG_ASSERT(BH((char *)b + b->bh.bb.bsize)->bb.bsize == ESent);
1021 KMP_DEBUG_ASSERT(BH((char *)b + b->bh.bb.bsize)->bb.prevfree ==
1022 b->bh.bb.bsize);
1023
1024 /* Unlink the buffer from the free list */
1025 __kmp_bget_remove_from_freelist(b);
1026
1027 KE_TRACE(10, ("%%%%%% FREE( %p )\n", (void *)b));
1028
1029 (*thr->relfcn)(b);
1030 thr->numprel++; /* Nr of expansion block releases */
1031 thr->numpblk--; /* Total number of blocks */
1032 KMP_DEBUG_ASSERT(thr->numpblk == thr->numpget - thr->numprel);
1033 }
1034 #endif /* BufStats */
1035
1036 /* Deallocate bget_data */
1037 if (th->th.th_local.bget_data != NULL) {
1038 __kmp_free(th->th.th_local.bget_data);
1039 th->th.th_local.bget_data = NULL;
1040 }
1041 }
1042
kmpc_set_poolsize(size_t size)1043 void kmpc_set_poolsize(size_t size) {
1044 bectl(__kmp_get_thread(), (bget_compact_t)0, (bget_acquire_t)malloc,
1045 (bget_release_t)free, (bufsize)size);
1046 }
1047
kmpc_get_poolsize(void)1048 size_t kmpc_get_poolsize(void) {
1049 thr_data_t *p;
1050
1051 p = get_thr_data(__kmp_get_thread());
1052
1053 return p->exp_incr;
1054 }
1055
kmpc_set_poolmode(int mode)1056 void kmpc_set_poolmode(int mode) {
1057 thr_data_t *p;
1058
1059 if (mode == bget_mode_fifo || mode == bget_mode_lifo ||
1060 mode == bget_mode_best) {
1061 p = get_thr_data(__kmp_get_thread());
1062 p->mode = (bget_mode_t)mode;
1063 }
1064 }
1065
kmpc_get_poolmode(void)1066 int kmpc_get_poolmode(void) {
1067 thr_data_t *p;
1068
1069 p = get_thr_data(__kmp_get_thread());
1070
1071 return p->mode;
1072 }
1073
kmpc_get_poolstat(size_t * maxmem,size_t * allmem)1074 void kmpc_get_poolstat(size_t *maxmem, size_t *allmem) {
1075 kmp_info_t *th = __kmp_get_thread();
1076 bufsize a, b;
1077
1078 __kmp_bget_dequeue(th); /* Release any queued buffers */
1079
1080 bcheck(th, &a, &b);
1081
1082 *maxmem = a;
1083 *allmem = b;
1084 }
1085
kmpc_poolprint(void)1086 void kmpc_poolprint(void) {
1087 kmp_info_t *th = __kmp_get_thread();
1088
1089 __kmp_bget_dequeue(th); /* Release any queued buffers */
1090
1091 bfreed(th);
1092 }
1093
1094 #endif // #if KMP_USE_BGET
1095
kmpc_malloc(size_t size)1096 void *kmpc_malloc(size_t size) {
1097 void *ptr;
1098 ptr = bget(__kmp_entry_thread(), (bufsize)(size + sizeof(ptr)));
1099 if (ptr != NULL) {
1100 // save allocated pointer just before one returned to user
1101 *(void **)ptr = ptr;
1102 ptr = (void **)ptr + 1;
1103 }
1104 return ptr;
1105 }
1106
1107 #define IS_POWER_OF_TWO(n) (((n) & ((n)-1)) == 0)
1108
kmpc_aligned_malloc(size_t size,size_t alignment)1109 void *kmpc_aligned_malloc(size_t size, size_t alignment) {
1110 void *ptr;
1111 void *ptr_allocated;
1112 KMP_DEBUG_ASSERT(alignment < 32 * 1024); // Alignment should not be too big
1113 if (!IS_POWER_OF_TWO(alignment)) {
1114 // AC: do we need to issue a warning here?
1115 errno = EINVAL;
1116 return NULL;
1117 }
1118 size = size + sizeof(void *) + alignment;
1119 ptr_allocated = bget(__kmp_entry_thread(), (bufsize)size);
1120 if (ptr_allocated != NULL) {
1121 // save allocated pointer just before one returned to user
1122 ptr = (void *)(((kmp_uintptr_t)ptr_allocated + sizeof(void *) + alignment) &
1123 ~(alignment - 1));
1124 *((void **)ptr - 1) = ptr_allocated;
1125 } else {
1126 ptr = NULL;
1127 }
1128 return ptr;
1129 }
1130
kmpc_calloc(size_t nelem,size_t elsize)1131 void *kmpc_calloc(size_t nelem, size_t elsize) {
1132 void *ptr;
1133 ptr = bgetz(__kmp_entry_thread(), (bufsize)(nelem * elsize + sizeof(ptr)));
1134 if (ptr != NULL) {
1135 // save allocated pointer just before one returned to user
1136 *(void **)ptr = ptr;
1137 ptr = (void **)ptr + 1;
1138 }
1139 return ptr;
1140 }
1141
kmpc_realloc(void * ptr,size_t size)1142 void *kmpc_realloc(void *ptr, size_t size) {
1143 void *result = NULL;
1144 if (ptr == NULL) {
1145 // If pointer is NULL, realloc behaves like malloc.
1146 result = bget(__kmp_entry_thread(), (bufsize)(size + sizeof(ptr)));
1147 // save allocated pointer just before one returned to user
1148 if (result != NULL) {
1149 *(void **)result = result;
1150 result = (void **)result + 1;
1151 }
1152 } else if (size == 0) {
1153 // If size is 0, realloc behaves like free.
1154 // The thread must be registered by the call to kmpc_malloc() or
1155 // kmpc_calloc() before.
1156 // So it should be safe to call __kmp_get_thread(), not
1157 // __kmp_entry_thread().
1158 KMP_ASSERT(*((void **)ptr - 1));
1159 brel(__kmp_get_thread(), *((void **)ptr - 1));
1160 } else {
1161 result = bgetr(__kmp_entry_thread(), *((void **)ptr - 1),
1162 (bufsize)(size + sizeof(ptr)));
1163 if (result != NULL) {
1164 *(void **)result = result;
1165 result = (void **)result + 1;
1166 }
1167 }
1168 return result;
1169 }
1170
1171 // NOTE: the library must have already been initialized by a previous allocate
kmpc_free(void * ptr)1172 void kmpc_free(void *ptr) {
1173 if (!__kmp_init_serial) {
1174 return;
1175 }
1176 if (ptr != NULL) {
1177 kmp_info_t *th = __kmp_get_thread();
1178 __kmp_bget_dequeue(th); /* Release any queued buffers */
1179 // extract allocated pointer and free it
1180 KMP_ASSERT(*((void **)ptr - 1));
1181 brel(th, *((void **)ptr - 1));
1182 }
1183 }
1184
___kmp_thread_malloc(kmp_info_t * th,size_t size KMP_SRC_LOC_DECL)1185 void *___kmp_thread_malloc(kmp_info_t *th, size_t size KMP_SRC_LOC_DECL) {
1186 void *ptr;
1187 KE_TRACE(30, ("-> __kmp_thread_malloc( %p, %d ) called from %s:%d\n", th,
1188 (int)size KMP_SRC_LOC_PARM));
1189 ptr = bget(th, (bufsize)size);
1190 KE_TRACE(30, ("<- __kmp_thread_malloc() returns %p\n", ptr));
1191 return ptr;
1192 }
1193
___kmp_thread_calloc(kmp_info_t * th,size_t nelem,size_t elsize KMP_SRC_LOC_DECL)1194 void *___kmp_thread_calloc(kmp_info_t *th, size_t nelem,
1195 size_t elsize KMP_SRC_LOC_DECL) {
1196 void *ptr;
1197 KE_TRACE(30, ("-> __kmp_thread_calloc( %p, %d, %d ) called from %s:%d\n", th,
1198 (int)nelem, (int)elsize KMP_SRC_LOC_PARM));
1199 ptr = bgetz(th, (bufsize)(nelem * elsize));
1200 KE_TRACE(30, ("<- __kmp_thread_calloc() returns %p\n", ptr));
1201 return ptr;
1202 }
1203
___kmp_thread_realloc(kmp_info_t * th,void * ptr,size_t size KMP_SRC_LOC_DECL)1204 void *___kmp_thread_realloc(kmp_info_t *th, void *ptr,
1205 size_t size KMP_SRC_LOC_DECL) {
1206 KE_TRACE(30, ("-> __kmp_thread_realloc( %p, %p, %d ) called from %s:%d\n", th,
1207 ptr, (int)size KMP_SRC_LOC_PARM));
1208 ptr = bgetr(th, ptr, (bufsize)size);
1209 KE_TRACE(30, ("<- __kmp_thread_realloc() returns %p\n", ptr));
1210 return ptr;
1211 }
1212
___kmp_thread_free(kmp_info_t * th,void * ptr KMP_SRC_LOC_DECL)1213 void ___kmp_thread_free(kmp_info_t *th, void *ptr KMP_SRC_LOC_DECL) {
1214 KE_TRACE(30, ("-> __kmp_thread_free( %p, %p ) called from %s:%d\n", th,
1215 ptr KMP_SRC_LOC_PARM));
1216 if (ptr != NULL) {
1217 __kmp_bget_dequeue(th); /* Release any queued buffers */
1218 brel(th, ptr);
1219 }
1220 KE_TRACE(30, ("<- __kmp_thread_free()\n"));
1221 }
1222
1223 /* OMP 5.0 Memory Management support */
1224 static const char *kmp_mk_lib_name;
1225 static void *h_memkind;
1226 /* memkind experimental API: */
1227 // memkind_alloc
1228 static void *(*kmp_mk_alloc)(void *k, size_t sz);
1229 // memkind_free
1230 static void (*kmp_mk_free)(void *kind, void *ptr);
1231 // memkind_check_available
1232 static int (*kmp_mk_check)(void *kind);
1233 // kinds we are going to use
1234 static void **mk_default;
1235 static void **mk_interleave;
1236 static void **mk_hbw;
1237 static void **mk_hbw_interleave;
1238 static void **mk_hbw_preferred;
1239 static void **mk_hugetlb;
1240 static void **mk_hbw_hugetlb;
1241 static void **mk_hbw_preferred_hugetlb;
1242
1243 #if KMP_OS_UNIX && KMP_DYNAMIC_LIB
chk_kind(void *** pkind)1244 static inline void chk_kind(void ***pkind) {
1245 KMP_DEBUG_ASSERT(pkind);
1246 if (*pkind) // symbol found
1247 if (kmp_mk_check(**pkind)) // kind not available or error
1248 *pkind = NULL;
1249 }
1250 #endif
1251
__kmp_init_memkind()1252 void __kmp_init_memkind() {
1253 // as of 2018-07-31 memkind does not support Windows*, exclude it for now
1254 #if KMP_OS_UNIX && KMP_DYNAMIC_LIB
1255 // use of statically linked memkind is problematic, as it depends on libnuma
1256 kmp_mk_lib_name = "libmemkind.so";
1257 h_memkind = dlopen(kmp_mk_lib_name, RTLD_LAZY);
1258 if (h_memkind) {
1259 kmp_mk_check = (int (*)(void *))dlsym(h_memkind, "memkind_check_available");
1260 kmp_mk_alloc =
1261 (void *(*)(void *, size_t))dlsym(h_memkind, "memkind_malloc");
1262 kmp_mk_free = (void (*)(void *, void *))dlsym(h_memkind, "memkind_free");
1263 mk_default = (void **)dlsym(h_memkind, "MEMKIND_DEFAULT");
1264 if (kmp_mk_check && kmp_mk_alloc && kmp_mk_free && mk_default &&
1265 !kmp_mk_check(*mk_default)) {
1266 __kmp_memkind_available = 1;
1267 mk_interleave = (void **)dlsym(h_memkind, "MEMKIND_INTERLEAVE");
1268 chk_kind(&mk_interleave);
1269 mk_hbw = (void **)dlsym(h_memkind, "MEMKIND_HBW");
1270 chk_kind(&mk_hbw);
1271 mk_hbw_interleave = (void **)dlsym(h_memkind, "MEMKIND_HBW_INTERLEAVE");
1272 chk_kind(&mk_hbw_interleave);
1273 mk_hbw_preferred = (void **)dlsym(h_memkind, "MEMKIND_HBW_PREFERRED");
1274 chk_kind(&mk_hbw_preferred);
1275 mk_hugetlb = (void **)dlsym(h_memkind, "MEMKIND_HUGETLB");
1276 chk_kind(&mk_hugetlb);
1277 mk_hbw_hugetlb = (void **)dlsym(h_memkind, "MEMKIND_HBW_HUGETLB");
1278 chk_kind(&mk_hbw_hugetlb);
1279 mk_hbw_preferred_hugetlb =
1280 (void **)dlsym(h_memkind, "MEMKIND_HBW_PREFERRED_HUGETLB");
1281 chk_kind(&mk_hbw_preferred_hugetlb);
1282 KE_TRACE(25, ("__kmp_init_memkind: memkind library initialized\n"));
1283 return; // success
1284 }
1285 dlclose(h_memkind); // failure
1286 h_memkind = NULL;
1287 }
1288 kmp_mk_check = NULL;
1289 kmp_mk_alloc = NULL;
1290 kmp_mk_free = NULL;
1291 mk_default = NULL;
1292 mk_interleave = NULL;
1293 mk_hbw = NULL;
1294 mk_hbw_interleave = NULL;
1295 mk_hbw_preferred = NULL;
1296 mk_hugetlb = NULL;
1297 mk_hbw_hugetlb = NULL;
1298 mk_hbw_preferred_hugetlb = NULL;
1299 #else
1300 kmp_mk_lib_name = "";
1301 h_memkind = NULL;
1302 kmp_mk_check = NULL;
1303 kmp_mk_alloc = NULL;
1304 kmp_mk_free = NULL;
1305 mk_default = NULL;
1306 mk_interleave = NULL;
1307 mk_hbw = NULL;
1308 mk_hbw_interleave = NULL;
1309 mk_hbw_preferred = NULL;
1310 mk_hugetlb = NULL;
1311 mk_hbw_hugetlb = NULL;
1312 mk_hbw_preferred_hugetlb = NULL;
1313 #endif
1314 }
1315
__kmp_fini_memkind()1316 void __kmp_fini_memkind() {
1317 #if KMP_OS_UNIX && KMP_DYNAMIC_LIB
1318 if (__kmp_memkind_available)
1319 KE_TRACE(25, ("__kmp_fini_memkind: finalize memkind library\n"));
1320 if (h_memkind) {
1321 dlclose(h_memkind);
1322 h_memkind = NULL;
1323 }
1324 kmp_mk_check = NULL;
1325 kmp_mk_alloc = NULL;
1326 kmp_mk_free = NULL;
1327 mk_default = NULL;
1328 mk_interleave = NULL;
1329 mk_hbw = NULL;
1330 mk_hbw_interleave = NULL;
1331 mk_hbw_preferred = NULL;
1332 mk_hugetlb = NULL;
1333 mk_hbw_hugetlb = NULL;
1334 mk_hbw_preferred_hugetlb = NULL;
1335 #endif
1336 }
1337
__kmpc_init_allocator(int gtid,omp_memspace_handle_t ms,int ntraits,omp_alloctrait_t traits[])1338 omp_allocator_handle_t __kmpc_init_allocator(int gtid, omp_memspace_handle_t ms,
1339 int ntraits,
1340 omp_alloctrait_t traits[]) {
1341 // OpenMP 5.0 only allows predefined memspaces
1342 KMP_DEBUG_ASSERT(ms == omp_default_mem_space || ms == omp_low_lat_mem_space ||
1343 ms == omp_large_cap_mem_space || ms == omp_const_mem_space ||
1344 ms == omp_high_bw_mem_space);
1345 kmp_allocator_t *al;
1346 int i;
1347 al = (kmp_allocator_t *)__kmp_allocate(sizeof(kmp_allocator_t)); // zeroed
1348 al->memspace = ms; // not used currently
1349 for (i = 0; i < ntraits; ++i) {
1350 switch (traits[i].key) {
1351 case omp_atk_threadmodel:
1352 case omp_atk_access:
1353 case omp_atk_pinned:
1354 break;
1355 case omp_atk_alignment:
1356 al->alignment = traits[i].value;
1357 KMP_ASSERT(IS_POWER_OF_TWO(al->alignment));
1358 break;
1359 case omp_atk_pool_size:
1360 al->pool_size = traits[i].value;
1361 break;
1362 case omp_atk_fallback:
1363 al->fb = (omp_alloctrait_value_t)traits[i].value;
1364 KMP_DEBUG_ASSERT(
1365 al->fb == omp_atv_default_mem_fb || al->fb == omp_atv_null_fb ||
1366 al->fb == omp_atv_abort_fb || al->fb == omp_atv_allocator_fb);
1367 break;
1368 case omp_atk_fb_data:
1369 al->fb_data = RCAST(kmp_allocator_t *, traits[i].value);
1370 break;
1371 case omp_atk_partition:
1372 al->memkind = RCAST(void **, traits[i].value);
1373 break;
1374 default:
1375 KMP_ASSERT2(0, "Unexpected allocator trait");
1376 }
1377 }
1378 if (al->fb == 0) {
1379 // set default allocator
1380 al->fb = omp_atv_default_mem_fb;
1381 al->fb_data = (kmp_allocator_t *)omp_default_mem_alloc;
1382 } else if (al->fb == omp_atv_allocator_fb) {
1383 KMP_ASSERT(al->fb_data != NULL);
1384 } else if (al->fb == omp_atv_default_mem_fb) {
1385 al->fb_data = (kmp_allocator_t *)omp_default_mem_alloc;
1386 }
1387 if (__kmp_memkind_available) {
1388 // Let's use memkind library if available
1389 if (ms == omp_high_bw_mem_space) {
1390 if (al->memkind == (void *)omp_atv_interleaved && mk_hbw_interleave) {
1391 al->memkind = mk_hbw_interleave;
1392 } else if (mk_hbw_preferred) {
1393 // AC: do not try to use MEMKIND_HBW for now, because memkind library
1394 // cannot reliably detect exhaustion of HBW memory.
1395 // It could be possible using hbw_verify_memory_region() but memkind
1396 // manual says: "Using this function in production code may result in
1397 // serious performance penalty".
1398 al->memkind = mk_hbw_preferred;
1399 } else {
1400 // HBW is requested but not available --> return NULL allocator
1401 __kmp_free(al);
1402 return omp_null_allocator;
1403 }
1404 } else {
1405 if (al->memkind == (void *)omp_atv_interleaved && mk_interleave) {
1406 al->memkind = mk_interleave;
1407 } else {
1408 al->memkind = mk_default;
1409 }
1410 }
1411 } else {
1412 if (ms == omp_high_bw_mem_space) {
1413 // cannot detect HBW memory presence without memkind library
1414 __kmp_free(al);
1415 return omp_null_allocator;
1416 }
1417 }
1418 return (omp_allocator_handle_t)al;
1419 }
1420
__kmpc_destroy_allocator(int gtid,omp_allocator_handle_t allocator)1421 void __kmpc_destroy_allocator(int gtid, omp_allocator_handle_t allocator) {
1422 if (allocator > kmp_max_mem_alloc)
1423 __kmp_free(allocator);
1424 }
1425
__kmpc_set_default_allocator(int gtid,omp_allocator_handle_t allocator)1426 void __kmpc_set_default_allocator(int gtid, omp_allocator_handle_t allocator) {
1427 if (allocator == omp_null_allocator)
1428 allocator = omp_default_mem_alloc;
1429 __kmp_threads[gtid]->th.th_def_allocator = allocator;
1430 }
1431
__kmpc_get_default_allocator(int gtid)1432 omp_allocator_handle_t __kmpc_get_default_allocator(int gtid) {
1433 return __kmp_threads[gtid]->th.th_def_allocator;
1434 }
1435
1436 typedef struct kmp_mem_desc { // Memory block descriptor
1437 void *ptr_alloc; // Pointer returned by allocator
1438 size_t size_a; // Size of allocated memory block (initial+descriptor+align)
1439 size_t size_orig; // Original size requested
1440 void *ptr_align; // Pointer to aligned memory, returned
1441 kmp_allocator_t *allocator; // allocator
1442 } kmp_mem_desc_t;
1443 static int alignment = sizeof(void *); // let's align to pointer size
1444
__kmpc_alloc(int gtid,size_t size,omp_allocator_handle_t allocator)1445 void *__kmpc_alloc(int gtid, size_t size, omp_allocator_handle_t allocator) {
1446 void *ptr = NULL;
1447 kmp_allocator_t *al;
1448 KMP_DEBUG_ASSERT(__kmp_init_serial);
1449
1450 if (size == 0)
1451 return NULL;
1452
1453 if (allocator == omp_null_allocator)
1454 allocator = __kmp_threads[gtid]->th.th_def_allocator;
1455
1456 KE_TRACE(25, ("__kmpc_alloc: T#%d (%d, %p)\n", gtid, (int)size, allocator));
1457 al = RCAST(kmp_allocator_t *, CCAST(omp_allocator_handle_t, allocator));
1458
1459 int sz_desc = sizeof(kmp_mem_desc_t);
1460 kmp_mem_desc_t desc;
1461 kmp_uintptr_t addr; // address returned by allocator
1462 kmp_uintptr_t addr_align; // address to return to caller
1463 kmp_uintptr_t addr_descr; // address of memory block descriptor
1464 int align = alignment; // default alignment
1465 if (allocator > kmp_max_mem_alloc && al->alignment > 0) {
1466 align = al->alignment; // alignment requested by user
1467 }
1468 desc.size_orig = size;
1469 desc.size_a = size + sz_desc + align;
1470
1471 if (__kmp_memkind_available) {
1472 if (allocator < kmp_max_mem_alloc) {
1473 // pre-defined allocator
1474 if (allocator == omp_high_bw_mem_alloc && mk_hbw_preferred) {
1475 ptr = kmp_mk_alloc(*mk_hbw_preferred, desc.size_a);
1476 } else {
1477 ptr = kmp_mk_alloc(*mk_default, desc.size_a);
1478 }
1479 } else if (al->pool_size > 0) {
1480 // custom allocator with pool size requested
1481 kmp_uint64 used =
1482 KMP_TEST_THEN_ADD64((kmp_int64 *)&al->pool_used, desc.size_a);
1483 if (used + desc.size_a > al->pool_size) {
1484 // not enough space, need to go fallback path
1485 KMP_TEST_THEN_ADD64((kmp_int64 *)&al->pool_used, -desc.size_a);
1486 if (al->fb == omp_atv_default_mem_fb) {
1487 al = (kmp_allocator_t *)omp_default_mem_alloc;
1488 ptr = kmp_mk_alloc(*mk_default, desc.size_a);
1489 } else if (al->fb == omp_atv_abort_fb) {
1490 KMP_ASSERT(0); // abort fallback requested
1491 } else if (al->fb == omp_atv_allocator_fb) {
1492 KMP_ASSERT(al != al->fb_data);
1493 al = al->fb_data;
1494 return __kmpc_alloc(gtid, size, (omp_allocator_handle_t)al);
1495 } // else ptr == NULL;
1496 } else {
1497 // pool has enough space
1498 ptr = kmp_mk_alloc(*al->memkind, desc.size_a);
1499 if (ptr == NULL) {
1500 if (al->fb == omp_atv_default_mem_fb) {
1501 al = (kmp_allocator_t *)omp_default_mem_alloc;
1502 ptr = kmp_mk_alloc(*mk_default, desc.size_a);
1503 } else if (al->fb == omp_atv_abort_fb) {
1504 KMP_ASSERT(0); // abort fallback requested
1505 } else if (al->fb == omp_atv_allocator_fb) {
1506 KMP_ASSERT(al != al->fb_data);
1507 al = al->fb_data;
1508 return __kmpc_alloc(gtid, size, (omp_allocator_handle_t)al);
1509 }
1510 }
1511 }
1512 } else {
1513 // custom allocator, pool size not requested
1514 ptr = kmp_mk_alloc(*al->memkind, desc.size_a);
1515 if (ptr == NULL) {
1516 if (al->fb == omp_atv_default_mem_fb) {
1517 al = (kmp_allocator_t *)omp_default_mem_alloc;
1518 ptr = kmp_mk_alloc(*mk_default, desc.size_a);
1519 } else if (al->fb == omp_atv_abort_fb) {
1520 KMP_ASSERT(0); // abort fallback requested
1521 } else if (al->fb == omp_atv_allocator_fb) {
1522 KMP_ASSERT(al != al->fb_data);
1523 al = al->fb_data;
1524 return __kmpc_alloc(gtid, size, (omp_allocator_handle_t)al);
1525 }
1526 }
1527 }
1528 } else if (allocator < kmp_max_mem_alloc) {
1529 // pre-defined allocator
1530 if (allocator == omp_high_bw_mem_alloc) {
1531 // ptr = NULL;
1532 } else {
1533 ptr = __kmp_thread_malloc(__kmp_thread_from_gtid(gtid), desc.size_a);
1534 }
1535 } else if (al->pool_size > 0) {
1536 // custom allocator with pool size requested
1537 kmp_uint64 used =
1538 KMP_TEST_THEN_ADD64((kmp_int64 *)&al->pool_used, desc.size_a);
1539 if (used + desc.size_a > al->pool_size) {
1540 // not enough space, need to go fallback path
1541 KMP_TEST_THEN_ADD64((kmp_int64 *)&al->pool_used, -desc.size_a);
1542 if (al->fb == omp_atv_default_mem_fb) {
1543 al = (kmp_allocator_t *)omp_default_mem_alloc;
1544 ptr = __kmp_thread_malloc(__kmp_thread_from_gtid(gtid), desc.size_a);
1545 } else if (al->fb == omp_atv_abort_fb) {
1546 KMP_ASSERT(0); // abort fallback requested
1547 } else if (al->fb == omp_atv_allocator_fb) {
1548 KMP_ASSERT(al != al->fb_data);
1549 al = al->fb_data;
1550 return __kmpc_alloc(gtid, size, (omp_allocator_handle_t)al);
1551 } // else ptr == NULL;
1552 } else {
1553 // pool has enough space
1554 ptr = __kmp_thread_malloc(__kmp_thread_from_gtid(gtid), desc.size_a);
1555 if (ptr == NULL && al->fb == omp_atv_abort_fb) {
1556 KMP_ASSERT(0); // abort fallback requested
1557 } // no sense to look for another fallback because of same internal alloc
1558 }
1559 } else {
1560 // custom allocator, pool size not requested
1561 ptr = __kmp_thread_malloc(__kmp_thread_from_gtid(gtid), desc.size_a);
1562 if (ptr == NULL && al->fb == omp_atv_abort_fb) {
1563 KMP_ASSERT(0); // abort fallback requested
1564 } // no sense to look for another fallback because of same internal alloc
1565 }
1566 KE_TRACE(10, ("__kmpc_alloc: T#%d %p=alloc(%d)\n", gtid, ptr, desc.size_a));
1567 if (ptr == NULL)
1568 return NULL;
1569
1570 addr = (kmp_uintptr_t)ptr;
1571 addr_align = (addr + sz_desc + align - 1) & ~(align - 1);
1572 addr_descr = addr_align - sz_desc;
1573
1574 desc.ptr_alloc = ptr;
1575 desc.ptr_align = (void *)addr_align;
1576 desc.allocator = al;
1577 *((kmp_mem_desc_t *)addr_descr) = desc; // save descriptor contents
1578 KMP_MB();
1579
1580 KE_TRACE(25, ("__kmpc_alloc returns %p, T#%d\n", desc.ptr_align, gtid));
1581 return desc.ptr_align;
1582 }
1583
__kmpc_calloc(int gtid,size_t nmemb,size_t size,omp_allocator_handle_t allocator)1584 void *__kmpc_calloc(int gtid, size_t nmemb, size_t size,
1585 omp_allocator_handle_t allocator) {
1586 void *ptr = NULL;
1587 kmp_allocator_t *al;
1588 KMP_DEBUG_ASSERT(__kmp_init_serial);
1589
1590 if (allocator == omp_null_allocator)
1591 allocator = __kmp_threads[gtid]->th.th_def_allocator;
1592
1593 KE_TRACE(25, ("__kmpc_calloc: T#%d (%d, %d, %p)\n", gtid, (int)nmemb,
1594 (int)size, allocator));
1595
1596 al = RCAST(kmp_allocator_t *, CCAST(omp_allocator_handle_t, allocator));
1597
1598 if (nmemb == 0 || size == 0)
1599 return ptr;
1600
1601 if ((SIZE_MAX - sizeof(kmp_mem_desc_t)) / size < nmemb) {
1602 if (al->fb == omp_atv_abort_fb) {
1603 KMP_ASSERT(0);
1604 }
1605 return ptr;
1606 }
1607
1608 ptr = __kmpc_alloc(gtid, nmemb * size, allocator);
1609
1610 if (ptr) {
1611 memset(ptr, 0x00, nmemb * size);
1612 }
1613 KE_TRACE(25, ("__kmpc_calloc returns %p, T#%d\n", ptr, gtid));
1614 return ptr;
1615 }
1616
__kmpc_realloc(int gtid,void * ptr,size_t size,omp_allocator_handle_t allocator,omp_allocator_handle_t free_allocator)1617 void *__kmpc_realloc(int gtid, void *ptr, size_t size,
1618 omp_allocator_handle_t allocator,
1619 omp_allocator_handle_t free_allocator) {
1620 void *nptr = NULL;
1621 KMP_DEBUG_ASSERT(__kmp_init_serial);
1622
1623 if (size == 0) {
1624 if (ptr != NULL)
1625 __kmpc_free(gtid, ptr, free_allocator);
1626 return nptr;
1627 }
1628
1629 KE_TRACE(25, ("__kmpc_realloc: T#%d (%p, %d, %p, %p)\n", gtid, ptr, (int)size,
1630 allocator, free_allocator));
1631
1632 nptr = __kmpc_alloc(gtid, size, allocator);
1633
1634 if (nptr != NULL && ptr != NULL) {
1635 kmp_mem_desc_t desc;
1636 kmp_uintptr_t addr_align; // address to return to caller
1637 kmp_uintptr_t addr_descr; // address of memory block descriptor
1638
1639 addr_align = (kmp_uintptr_t)ptr;
1640 addr_descr = addr_align - sizeof(kmp_mem_desc_t);
1641 desc = *((kmp_mem_desc_t *)addr_descr); // read descriptor
1642
1643 KMP_DEBUG_ASSERT(desc.ptr_align == ptr);
1644 KMP_DEBUG_ASSERT(desc.size_orig > 0);
1645 KMP_DEBUG_ASSERT(desc.size_orig < desc.size_a);
1646 KMP_MEMCPY((char *)nptr, (char *)ptr,
1647 (size_t)((size < desc.size_orig) ? size : desc.size_orig));
1648 }
1649
1650 if (nptr != NULL) {
1651 __kmpc_free(gtid, ptr, free_allocator);
1652 }
1653
1654 KE_TRACE(25, ("__kmpc_realloc returns %p, T#%d\n", nptr, gtid));
1655 return nptr;
1656 }
1657
__kmpc_free(int gtid,void * ptr,const omp_allocator_handle_t allocator)1658 void __kmpc_free(int gtid, void *ptr, const omp_allocator_handle_t allocator) {
1659 KE_TRACE(25, ("__kmpc_free: T#%d free(%p,%p)\n", gtid, ptr, allocator));
1660 if (ptr == NULL)
1661 return;
1662
1663 kmp_allocator_t *al;
1664 omp_allocator_handle_t oal;
1665 al = RCAST(kmp_allocator_t *, CCAST(omp_allocator_handle_t, allocator));
1666 kmp_mem_desc_t desc;
1667 kmp_uintptr_t addr_align; // address to return to caller
1668 kmp_uintptr_t addr_descr; // address of memory block descriptor
1669
1670 addr_align = (kmp_uintptr_t)ptr;
1671 addr_descr = addr_align - sizeof(kmp_mem_desc_t);
1672 desc = *((kmp_mem_desc_t *)addr_descr); // read descriptor
1673
1674 KMP_DEBUG_ASSERT(desc.ptr_align == ptr);
1675 if (allocator) {
1676 KMP_DEBUG_ASSERT(desc.allocator == al || desc.allocator == al->fb_data);
1677 }
1678 al = desc.allocator;
1679 oal = (omp_allocator_handle_t)al; // cast to void* for comparisons
1680 KMP_DEBUG_ASSERT(al);
1681
1682 if (__kmp_memkind_available) {
1683 if (oal < kmp_max_mem_alloc) {
1684 // pre-defined allocator
1685 if (oal == omp_high_bw_mem_alloc && mk_hbw_preferred) {
1686 kmp_mk_free(*mk_hbw_preferred, desc.ptr_alloc);
1687 } else {
1688 kmp_mk_free(*mk_default, desc.ptr_alloc);
1689 }
1690 } else {
1691 if (al->pool_size > 0) { // custom allocator with pool size requested
1692 kmp_uint64 used =
1693 KMP_TEST_THEN_ADD64((kmp_int64 *)&al->pool_used, -desc.size_a);
1694 (void)used; // to suppress compiler warning
1695 KMP_DEBUG_ASSERT(used >= desc.size_a);
1696 }
1697 kmp_mk_free(*al->memkind, desc.ptr_alloc);
1698 }
1699 } else {
1700 if (oal > kmp_max_mem_alloc && al->pool_size > 0) {
1701 kmp_uint64 used =
1702 KMP_TEST_THEN_ADD64((kmp_int64 *)&al->pool_used, -desc.size_a);
1703 (void)used; // to suppress compiler warning
1704 KMP_DEBUG_ASSERT(used >= desc.size_a);
1705 }
1706 __kmp_thread_free(__kmp_thread_from_gtid(gtid), desc.ptr_alloc);
1707 }
1708 KE_TRACE(10, ("__kmpc_free: T#%d freed %p (%p)\n", gtid, desc.ptr_alloc,
1709 allocator));
1710 }
1711
1712 /* If LEAK_MEMORY is defined, __kmp_free() will *not* free memory. It causes
1713 memory leaks, but it may be useful for debugging memory corruptions, used
1714 freed pointers, etc. */
1715 /* #define LEAK_MEMORY */
1716 struct kmp_mem_descr { // Memory block descriptor.
1717 void *ptr_allocated; // Pointer returned by malloc(), subject for free().
1718 size_t size_allocated; // Size of allocated memory block.
1719 void *ptr_aligned; // Pointer to aligned memory, to be used by client code.
1720 size_t size_aligned; // Size of aligned memory block.
1721 };
1722 typedef struct kmp_mem_descr kmp_mem_descr_t;
1723
1724 /* Allocate memory on requested boundary, fill allocated memory with 0x00.
1725 NULL is NEVER returned, __kmp_abort() is called in case of memory allocation
1726 error. Must use __kmp_free when freeing memory allocated by this routine! */
___kmp_allocate_align(size_t size,size_t alignment KMP_SRC_LOC_DECL)1727 static void *___kmp_allocate_align(size_t size,
1728 size_t alignment KMP_SRC_LOC_DECL) {
1729 /* __kmp_allocate() allocates (by call to malloc()) bigger memory block than
1730 requested to return properly aligned pointer. Original pointer returned
1731 by malloc() and size of allocated block is saved in descriptor just
1732 before the aligned pointer. This information used by __kmp_free() -- it
1733 has to pass to free() original pointer, not aligned one.
1734
1735 +---------+------------+-----------------------------------+---------+
1736 | padding | descriptor | aligned block | padding |
1737 +---------+------------+-----------------------------------+---------+
1738 ^ ^
1739 | |
1740 | +- Aligned pointer returned to caller
1741 +- Pointer returned by malloc()
1742
1743 Aligned block is filled with zeros, paddings are filled with 0xEF. */
1744
1745 kmp_mem_descr_t descr;
1746 kmp_uintptr_t addr_allocated; // Address returned by malloc().
1747 kmp_uintptr_t addr_aligned; // Aligned address to return to caller.
1748 kmp_uintptr_t addr_descr; // Address of memory block descriptor.
1749
1750 KE_TRACE(25, ("-> ___kmp_allocate_align( %d, %d ) called from %s:%d\n",
1751 (int)size, (int)alignment KMP_SRC_LOC_PARM));
1752
1753 KMP_DEBUG_ASSERT(alignment < 32 * 1024); // Alignment should not be too
1754 KMP_DEBUG_ASSERT(sizeof(void *) <= sizeof(kmp_uintptr_t));
1755 // Make sure kmp_uintptr_t is enough to store addresses.
1756
1757 descr.size_aligned = size;
1758 descr.size_allocated =
1759 descr.size_aligned + sizeof(kmp_mem_descr_t) + alignment;
1760
1761 #if KMP_DEBUG
1762 descr.ptr_allocated = _malloc_src_loc(descr.size_allocated, _file_, _line_);
1763 #else
1764 descr.ptr_allocated = malloc_src_loc(descr.size_allocated KMP_SRC_LOC_PARM);
1765 #endif
1766 KE_TRACE(10, (" malloc( %d ) returned %p\n", (int)descr.size_allocated,
1767 descr.ptr_allocated));
1768 if (descr.ptr_allocated == NULL) {
1769 KMP_FATAL(OutOfHeapMemory);
1770 }
1771
1772 addr_allocated = (kmp_uintptr_t)descr.ptr_allocated;
1773 addr_aligned =
1774 (addr_allocated + sizeof(kmp_mem_descr_t) + alignment) & ~(alignment - 1);
1775 addr_descr = addr_aligned - sizeof(kmp_mem_descr_t);
1776
1777 descr.ptr_aligned = (void *)addr_aligned;
1778
1779 KE_TRACE(26, (" ___kmp_allocate_align: "
1780 "ptr_allocated=%p, size_allocated=%d, "
1781 "ptr_aligned=%p, size_aligned=%d\n",
1782 descr.ptr_allocated, (int)descr.size_allocated,
1783 descr.ptr_aligned, (int)descr.size_aligned));
1784
1785 KMP_DEBUG_ASSERT(addr_allocated <= addr_descr);
1786 KMP_DEBUG_ASSERT(addr_descr + sizeof(kmp_mem_descr_t) == addr_aligned);
1787 KMP_DEBUG_ASSERT(addr_aligned + descr.size_aligned <=
1788 addr_allocated + descr.size_allocated);
1789 KMP_DEBUG_ASSERT(addr_aligned % alignment == 0);
1790 #ifdef KMP_DEBUG
1791 memset(descr.ptr_allocated, 0xEF, descr.size_allocated);
1792 // Fill allocated memory block with 0xEF.
1793 #endif
1794 memset(descr.ptr_aligned, 0x00, descr.size_aligned);
1795 // Fill the aligned memory block (which is intended for using by caller) with
1796 // 0x00. Do not
1797 // put this filling under KMP_DEBUG condition! Many callers expect zeroed
1798 // memory. (Padding
1799 // bytes remain filled with 0xEF in debugging library.)
1800 *((kmp_mem_descr_t *)addr_descr) = descr;
1801
1802 KMP_MB();
1803
1804 KE_TRACE(25, ("<- ___kmp_allocate_align() returns %p\n", descr.ptr_aligned));
1805 return descr.ptr_aligned;
1806 } // func ___kmp_allocate_align
1807
1808 /* Allocate memory on cache line boundary, fill allocated memory with 0x00.
1809 Do not call this func directly! Use __kmp_allocate macro instead.
1810 NULL is NEVER returned, __kmp_abort() is called in case of memory allocation
1811 error. Must use __kmp_free when freeing memory allocated by this routine! */
___kmp_allocate(size_t size KMP_SRC_LOC_DECL)1812 void *___kmp_allocate(size_t size KMP_SRC_LOC_DECL) {
1813 void *ptr;
1814 KE_TRACE(25, ("-> __kmp_allocate( %d ) called from %s:%d\n",
1815 (int)size KMP_SRC_LOC_PARM));
1816 ptr = ___kmp_allocate_align(size, __kmp_align_alloc KMP_SRC_LOC_PARM);
1817 KE_TRACE(25, ("<- __kmp_allocate() returns %p\n", ptr));
1818 return ptr;
1819 } // func ___kmp_allocate
1820
1821 /* Allocate memory on page boundary, fill allocated memory with 0x00.
1822 Does not call this func directly! Use __kmp_page_allocate macro instead.
1823 NULL is NEVER returned, __kmp_abort() is called in case of memory allocation
1824 error. Must use __kmp_free when freeing memory allocated by this routine! */
___kmp_page_allocate(size_t size KMP_SRC_LOC_DECL)1825 void *___kmp_page_allocate(size_t size KMP_SRC_LOC_DECL) {
1826 int page_size = 8 * 1024;
1827 void *ptr;
1828
1829 KE_TRACE(25, ("-> __kmp_page_allocate( %d ) called from %s:%d\n",
1830 (int)size KMP_SRC_LOC_PARM));
1831 ptr = ___kmp_allocate_align(size, page_size KMP_SRC_LOC_PARM);
1832 KE_TRACE(25, ("<- __kmp_page_allocate( %d ) returns %p\n", (int)size, ptr));
1833 return ptr;
1834 } // ___kmp_page_allocate
1835
1836 /* Free memory allocated by __kmp_allocate() and __kmp_page_allocate().
1837 In debug mode, fill the memory block with 0xEF before call to free(). */
___kmp_free(void * ptr KMP_SRC_LOC_DECL)1838 void ___kmp_free(void *ptr KMP_SRC_LOC_DECL) {
1839 kmp_mem_descr_t descr;
1840 kmp_uintptr_t addr_allocated; // Address returned by malloc().
1841 kmp_uintptr_t addr_aligned; // Aligned address passed by caller.
1842
1843 KE_TRACE(25,
1844 ("-> __kmp_free( %p ) called from %s:%d\n", ptr KMP_SRC_LOC_PARM));
1845 KMP_ASSERT(ptr != NULL);
1846
1847 descr = *(kmp_mem_descr_t *)((kmp_uintptr_t)ptr - sizeof(kmp_mem_descr_t));
1848
1849 KE_TRACE(26, (" __kmp_free: "
1850 "ptr_allocated=%p, size_allocated=%d, "
1851 "ptr_aligned=%p, size_aligned=%d\n",
1852 descr.ptr_allocated, (int)descr.size_allocated,
1853 descr.ptr_aligned, (int)descr.size_aligned));
1854
1855 addr_allocated = (kmp_uintptr_t)descr.ptr_allocated;
1856 addr_aligned = (kmp_uintptr_t)descr.ptr_aligned;
1857
1858 KMP_DEBUG_ASSERT(addr_aligned % CACHE_LINE == 0);
1859 KMP_DEBUG_ASSERT(descr.ptr_aligned == ptr);
1860 KMP_DEBUG_ASSERT(addr_allocated + sizeof(kmp_mem_descr_t) <= addr_aligned);
1861 KMP_DEBUG_ASSERT(descr.size_aligned < descr.size_allocated);
1862 KMP_DEBUG_ASSERT(addr_aligned + descr.size_aligned <=
1863 addr_allocated + descr.size_allocated);
1864
1865 #ifdef KMP_DEBUG
1866 memset(descr.ptr_allocated, 0xEF, descr.size_allocated);
1867 // Fill memory block with 0xEF, it helps catch using freed memory.
1868 #endif
1869
1870 #ifndef LEAK_MEMORY
1871 KE_TRACE(10, (" free( %p )\n", descr.ptr_allocated));
1872 #ifdef KMP_DEBUG
1873 _free_src_loc(descr.ptr_allocated, _file_, _line_);
1874 #else
1875 free_src_loc(descr.ptr_allocated KMP_SRC_LOC_PARM);
1876 #endif
1877 #endif
1878 KMP_MB();
1879 KE_TRACE(25, ("<- __kmp_free() returns\n"));
1880 } // func ___kmp_free
1881
1882 #if USE_FAST_MEMORY == 3
1883 // Allocate fast memory by first scanning the thread's free lists
1884 // If a chunk the right size exists, grab it off the free list.
1885 // Otherwise allocate normally using kmp_thread_malloc.
1886
1887 // AC: How to choose the limit? Just get 16 for now...
1888 #define KMP_FREE_LIST_LIMIT 16
1889
1890 // Always use 128 bytes for determining buckets for caching memory blocks
1891 #define DCACHE_LINE 128
1892
___kmp_fast_allocate(kmp_info_t * this_thr,size_t size KMP_SRC_LOC_DECL)1893 void *___kmp_fast_allocate(kmp_info_t *this_thr, size_t size KMP_SRC_LOC_DECL) {
1894 void *ptr;
1895 int num_lines;
1896 int idx;
1897 int index;
1898 void *alloc_ptr;
1899 size_t alloc_size;
1900 kmp_mem_descr_t *descr;
1901
1902 KE_TRACE(25, ("-> __kmp_fast_allocate( T#%d, %d ) called from %s:%d\n",
1903 __kmp_gtid_from_thread(this_thr), (int)size KMP_SRC_LOC_PARM));
1904
1905 num_lines = (size + DCACHE_LINE - 1) / DCACHE_LINE;
1906 idx = num_lines - 1;
1907 KMP_DEBUG_ASSERT(idx >= 0);
1908 if (idx < 2) {
1909 index = 0; // idx is [ 0, 1 ], use first free list
1910 num_lines = 2; // 1, 2 cache lines or less than cache line
1911 } else if ((idx >>= 2) == 0) {
1912 index = 1; // idx is [ 2, 3 ], use second free list
1913 num_lines = 4; // 3, 4 cache lines
1914 } else if ((idx >>= 2) == 0) {
1915 index = 2; // idx is [ 4, 15 ], use third free list
1916 num_lines = 16; // 5, 6, ..., 16 cache lines
1917 } else if ((idx >>= 2) == 0) {
1918 index = 3; // idx is [ 16, 63 ], use fourth free list
1919 num_lines = 64; // 17, 18, ..., 64 cache lines
1920 } else {
1921 goto alloc_call; // 65 or more cache lines ( > 8KB ), don't use free lists
1922 }
1923
1924 ptr = this_thr->th.th_free_lists[index].th_free_list_self;
1925 if (ptr != NULL) {
1926 // pop the head of no-sync free list
1927 this_thr->th.th_free_lists[index].th_free_list_self = *((void **)ptr);
1928 KMP_DEBUG_ASSERT(
1929 this_thr ==
1930 ((kmp_mem_descr_t *)((kmp_uintptr_t)ptr - sizeof(kmp_mem_descr_t)))
1931 ->ptr_aligned);
1932 goto end;
1933 }
1934 ptr = TCR_SYNC_PTR(this_thr->th.th_free_lists[index].th_free_list_sync);
1935 if (ptr != NULL) {
1936 // no-sync free list is empty, use sync free list (filled in by other
1937 // threads only)
1938 // pop the head of the sync free list, push NULL instead
1939 while (!KMP_COMPARE_AND_STORE_PTR(
1940 &this_thr->th.th_free_lists[index].th_free_list_sync, ptr, nullptr)) {
1941 KMP_CPU_PAUSE();
1942 ptr = TCR_SYNC_PTR(this_thr->th.th_free_lists[index].th_free_list_sync);
1943 }
1944 // push the rest of chain into no-sync free list (can be NULL if there was
1945 // the only block)
1946 this_thr->th.th_free_lists[index].th_free_list_self = *((void **)ptr);
1947 KMP_DEBUG_ASSERT(
1948 this_thr ==
1949 ((kmp_mem_descr_t *)((kmp_uintptr_t)ptr - sizeof(kmp_mem_descr_t)))
1950 ->ptr_aligned);
1951 goto end;
1952 }
1953
1954 alloc_call:
1955 // haven't found block in the free lists, thus allocate it
1956 size = num_lines * DCACHE_LINE;
1957
1958 alloc_size = size + sizeof(kmp_mem_descr_t) + DCACHE_LINE;
1959 KE_TRACE(25, ("__kmp_fast_allocate: T#%d Calling __kmp_thread_malloc with "
1960 "alloc_size %d\n",
1961 __kmp_gtid_from_thread(this_thr), alloc_size));
1962 alloc_ptr = bget(this_thr, (bufsize)alloc_size);
1963
1964 // align ptr to DCACHE_LINE
1965 ptr = (void *)((((kmp_uintptr_t)alloc_ptr) + sizeof(kmp_mem_descr_t) +
1966 DCACHE_LINE) &
1967 ~(DCACHE_LINE - 1));
1968 descr = (kmp_mem_descr_t *)(((kmp_uintptr_t)ptr) - sizeof(kmp_mem_descr_t));
1969
1970 descr->ptr_allocated = alloc_ptr; // remember allocated pointer
1971 // we don't need size_allocated
1972 descr->ptr_aligned = (void *)this_thr; // remember allocating thread
1973 // (it is already saved in bget buffer,
1974 // but we may want to use another allocator in future)
1975 descr->size_aligned = size;
1976
1977 end:
1978 KE_TRACE(25, ("<- __kmp_fast_allocate( T#%d ) returns %p\n",
1979 __kmp_gtid_from_thread(this_thr), ptr));
1980 return ptr;
1981 } // func __kmp_fast_allocate
1982
1983 // Free fast memory and place it on the thread's free list if it is of
1984 // the correct size.
___kmp_fast_free(kmp_info_t * this_thr,void * ptr KMP_SRC_LOC_DECL)1985 void ___kmp_fast_free(kmp_info_t *this_thr, void *ptr KMP_SRC_LOC_DECL) {
1986 kmp_mem_descr_t *descr;
1987 kmp_info_t *alloc_thr;
1988 size_t size;
1989 size_t idx;
1990 int index;
1991
1992 KE_TRACE(25, ("-> __kmp_fast_free( T#%d, %p ) called from %s:%d\n",
1993 __kmp_gtid_from_thread(this_thr), ptr KMP_SRC_LOC_PARM));
1994 KMP_ASSERT(ptr != NULL);
1995
1996 descr = (kmp_mem_descr_t *)(((kmp_uintptr_t)ptr) - sizeof(kmp_mem_descr_t));
1997
1998 KE_TRACE(26, (" __kmp_fast_free: size_aligned=%d\n",
1999 (int)descr->size_aligned));
2000
2001 size = descr->size_aligned; // 2, 4, 16, 64, 65, 66, ... cache lines
2002
2003 idx = DCACHE_LINE * 2; // 2 cache lines is minimal size of block
2004 if (idx == size) {
2005 index = 0; // 2 cache lines
2006 } else if ((idx <<= 1) == size) {
2007 index = 1; // 4 cache lines
2008 } else if ((idx <<= 2) == size) {
2009 index = 2; // 16 cache lines
2010 } else if ((idx <<= 2) == size) {
2011 index = 3; // 64 cache lines
2012 } else {
2013 KMP_DEBUG_ASSERT(size > DCACHE_LINE * 64);
2014 goto free_call; // 65 or more cache lines ( > 8KB )
2015 }
2016
2017 alloc_thr = (kmp_info_t *)descr->ptr_aligned; // get thread owning the block
2018 if (alloc_thr == this_thr) {
2019 // push block to self no-sync free list, linking previous head (LIFO)
2020 *((void **)ptr) = this_thr->th.th_free_lists[index].th_free_list_self;
2021 this_thr->th.th_free_lists[index].th_free_list_self = ptr;
2022 } else {
2023 void *head = this_thr->th.th_free_lists[index].th_free_list_other;
2024 if (head == NULL) {
2025 // Create new free list
2026 this_thr->th.th_free_lists[index].th_free_list_other = ptr;
2027 *((void **)ptr) = NULL; // mark the tail of the list
2028 descr->size_allocated = (size_t)1; // head of the list keeps its length
2029 } else {
2030 // need to check existed "other" list's owner thread and size of queue
2031 kmp_mem_descr_t *dsc =
2032 (kmp_mem_descr_t *)((char *)head - sizeof(kmp_mem_descr_t));
2033 // allocating thread, same for all queue nodes
2034 kmp_info_t *q_th = (kmp_info_t *)(dsc->ptr_aligned);
2035 size_t q_sz =
2036 dsc->size_allocated + 1; // new size in case we add current task
2037 if (q_th == alloc_thr && q_sz <= KMP_FREE_LIST_LIMIT) {
2038 // we can add current task to "other" list, no sync needed
2039 *((void **)ptr) = head;
2040 descr->size_allocated = q_sz;
2041 this_thr->th.th_free_lists[index].th_free_list_other = ptr;
2042 } else {
2043 // either queue blocks owner is changing or size limit exceeded
2044 // return old queue to allocating thread (q_th) synchronously,
2045 // and start new list for alloc_thr's tasks
2046 void *old_ptr;
2047 void *tail = head;
2048 void *next = *((void **)head);
2049 while (next != NULL) {
2050 KMP_DEBUG_ASSERT(
2051 // queue size should decrease by 1 each step through the list
2052 ((kmp_mem_descr_t *)((char *)next - sizeof(kmp_mem_descr_t)))
2053 ->size_allocated +
2054 1 ==
2055 ((kmp_mem_descr_t *)((char *)tail - sizeof(kmp_mem_descr_t)))
2056 ->size_allocated);
2057 tail = next; // remember tail node
2058 next = *((void **)next);
2059 }
2060 KMP_DEBUG_ASSERT(q_th != NULL);
2061 // push block to owner's sync free list
2062 old_ptr = TCR_PTR(q_th->th.th_free_lists[index].th_free_list_sync);
2063 /* the next pointer must be set before setting free_list to ptr to avoid
2064 exposing a broken list to other threads, even for an instant. */
2065 *((void **)tail) = old_ptr;
2066
2067 while (!KMP_COMPARE_AND_STORE_PTR(
2068 &q_th->th.th_free_lists[index].th_free_list_sync, old_ptr, head)) {
2069 KMP_CPU_PAUSE();
2070 old_ptr = TCR_PTR(q_th->th.th_free_lists[index].th_free_list_sync);
2071 *((void **)tail) = old_ptr;
2072 }
2073
2074 // start new list of not-selt tasks
2075 this_thr->th.th_free_lists[index].th_free_list_other = ptr;
2076 *((void **)ptr) = NULL;
2077 descr->size_allocated = (size_t)1; // head of queue keeps its length
2078 }
2079 }
2080 }
2081 goto end;
2082
2083 free_call:
2084 KE_TRACE(25, ("__kmp_fast_free: T#%d Calling __kmp_thread_free for size %d\n",
2085 __kmp_gtid_from_thread(this_thr), size));
2086 __kmp_bget_dequeue(this_thr); /* Release any queued buffers */
2087 brel(this_thr, descr->ptr_allocated);
2088
2089 end:
2090 KE_TRACE(25, ("<- __kmp_fast_free() returns\n"));
2091
2092 } // func __kmp_fast_free
2093
2094 // Initialize the thread free lists related to fast memory
2095 // Only do this when a thread is initially created.
__kmp_initialize_fast_memory(kmp_info_t * this_thr)2096 void __kmp_initialize_fast_memory(kmp_info_t *this_thr) {
2097 KE_TRACE(10, ("__kmp_initialize_fast_memory: Called from th %p\n", this_thr));
2098
2099 memset(this_thr->th.th_free_lists, 0, NUM_LISTS * sizeof(kmp_free_list_t));
2100 }
2101
2102 // Free the memory in the thread free lists related to fast memory
2103 // Only do this when a thread is being reaped (destroyed).
__kmp_free_fast_memory(kmp_info_t * th)2104 void __kmp_free_fast_memory(kmp_info_t *th) {
2105 // Suppose we use BGET underlying allocator, walk through its structures...
2106 int bin;
2107 thr_data_t *thr = get_thr_data(th);
2108 void **lst = NULL;
2109
2110 KE_TRACE(
2111 5, ("__kmp_free_fast_memory: Called T#%d\n", __kmp_gtid_from_thread(th)));
2112
2113 __kmp_bget_dequeue(th); // Release any queued buffers
2114
2115 // Dig through free lists and extract all allocated blocks
2116 for (bin = 0; bin < MAX_BGET_BINS; ++bin) {
2117 bfhead_t *b = thr->freelist[bin].ql.flink;
2118 while (b != &thr->freelist[bin]) {
2119 if ((kmp_uintptr_t)b->bh.bb.bthr & 1) { // the buffer is allocated address
2120 *((void **)b) =
2121 lst; // link the list (override bthr, but keep flink yet)
2122 lst = (void **)b; // push b into lst
2123 }
2124 b = b->ql.flink; // get next buffer
2125 }
2126 }
2127 while (lst != NULL) {
2128 void *next = *lst;
2129 KE_TRACE(10, ("__kmp_free_fast_memory: freeing %p, next=%p th %p (%d)\n",
2130 lst, next, th, __kmp_gtid_from_thread(th)));
2131 (*thr->relfcn)(lst);
2132 #if BufStats
2133 // count blocks to prevent problems in __kmp_finalize_bget()
2134 thr->numprel++; /* Nr of expansion block releases */
2135 thr->numpblk--; /* Total number of blocks */
2136 #endif
2137 lst = (void **)next;
2138 }
2139
2140 KE_TRACE(
2141 5, ("__kmp_free_fast_memory: Freed T#%d\n", __kmp_gtid_from_thread(th)));
2142 }
2143
2144 #endif // USE_FAST_MEMORY
2145