1 /* SPDX-License-Identifier: BSD-3-Clause
3 * Copyright (c) 2017-2018 Solarflare Communications Inc.
6 * This software was jointly developed between OKTET Labs (under contract
7 * for Solarflare) and Solarflare Communications, Inc.
14 #include <rte_errno.h>
16 #include <rte_mempool.h>
17 #include <rte_malloc.h>
20 * The general idea of the bucket mempool driver is as follows.
21 * We keep track of physically contiguous groups (buckets) of objects
22 * of a certain size. Every such a group has a counter that is
23 * incremented every time an object from that group is enqueued.
24 * Until the bucket is full, no objects from it are eligible for allocation.
25 * If a request is made to dequeue a multiply of bucket size, it is
26 * satisfied by returning the whole buckets, instead of separate objects.
30 struct bucket_header {
31 unsigned int lcore_id;
42 unsigned int header_size;
43 unsigned int total_elt_size;
44 unsigned int obj_per_bucket;
45 unsigned int bucket_stack_thresh;
46 uintptr_t bucket_page_mask;
47 struct rte_ring *shared_bucket_ring;
48 struct bucket_stack *buckets[RTE_MAX_LCORE];
50 * Multi-producer single-consumer ring to hold objects that are
51 * returned to the mempool at a different lcore than initially
54 struct rte_ring *adoption_buffer_rings[RTE_MAX_LCORE];
55 struct rte_ring *shared_orphan_ring;
56 struct rte_mempool *pool;
57 unsigned int bucket_mem_size;
60 static struct bucket_stack *
61 bucket_stack_create(const struct rte_mempool *mp, unsigned int n_elts)
63 struct bucket_stack *stack;
65 stack = rte_zmalloc_socket("bucket_stack",
66 sizeof(struct bucket_stack) +
67 n_elts * sizeof(void *),
72 stack->limit = n_elts;
79 bucket_stack_push(struct bucket_stack *stack, void *obj)
81 RTE_ASSERT(stack->top < stack->limit);
82 stack->objects[stack->top++] = obj;
86 bucket_stack_pop_unsafe(struct bucket_stack *stack)
88 RTE_ASSERT(stack->top > 0);
89 return stack->objects[--stack->top];
93 bucket_stack_pop(struct bucket_stack *stack)
97 return bucket_stack_pop_unsafe(stack);
101 bucket_enqueue_single(struct bucket_data *bd, void *obj)
104 uintptr_t addr = (uintptr_t)obj;
105 struct bucket_header *hdr;
106 unsigned int lcore_id = rte_lcore_id();
108 addr &= bd->bucket_page_mask;
109 hdr = (struct bucket_header *)addr;
111 if (likely(hdr->lcore_id == lcore_id)) {
112 if (hdr->fill_cnt < bd->obj_per_bucket - 1) {
116 /* Stack is big enough to put all buckets */
117 bucket_stack_push(bd->buckets[lcore_id], hdr);
119 } else if (hdr->lcore_id != LCORE_ID_ANY) {
120 struct rte_ring *adopt_ring =
121 bd->adoption_buffer_rings[hdr->lcore_id];
123 rc = rte_ring_enqueue(adopt_ring, obj);
124 /* Ring is big enough to put all objects */
126 } else if (hdr->fill_cnt < bd->obj_per_bucket - 1) {
130 rc = rte_ring_enqueue(bd->shared_bucket_ring, hdr);
131 /* Ring is big enough to put all buckets */
139 bucket_enqueue(struct rte_mempool *mp, void * const *obj_table,
142 struct bucket_data *bd = mp->pool_data;
143 struct bucket_stack *local_stack = bd->buckets[rte_lcore_id()];
147 for (i = 0; i < n; i++) {
148 rc = bucket_enqueue_single(bd, obj_table[i]);
151 if (local_stack->top > bd->bucket_stack_thresh) {
152 rte_ring_enqueue_bulk(bd->shared_bucket_ring,
153 &local_stack->objects
154 [bd->bucket_stack_thresh],
156 bd->bucket_stack_thresh,
158 local_stack->top = bd->bucket_stack_thresh;
164 bucket_fill_obj_table(const struct bucket_data *bd, void **pstart,
165 void **obj_table, unsigned int n)
168 uint8_t *objptr = *pstart;
170 for (objptr += bd->header_size, i = 0; i < n;
171 i++, objptr += bd->total_elt_size)
172 *obj_table++ = objptr;
178 bucket_dequeue_orphans(struct bucket_data *bd, void **obj_table,
179 unsigned int n_orphans)
185 rc = rte_ring_dequeue_bulk(bd->shared_orphan_ring, obj_table,
187 if (unlikely(rc != (int)n_orphans)) {
188 struct bucket_header *hdr;
190 objptr = bucket_stack_pop(bd->buckets[rte_lcore_id()]);
191 hdr = (struct bucket_header *)objptr;
193 if (objptr == NULL) {
194 rc = rte_ring_dequeue(bd->shared_bucket_ring,
200 hdr = (struct bucket_header *)objptr;
201 hdr->lcore_id = rte_lcore_id();
204 bucket_fill_obj_table(bd, (void **)&objptr, obj_table,
206 for (i = n_orphans; i < bd->obj_per_bucket; i++,
207 objptr += bd->total_elt_size) {
208 rc = rte_ring_enqueue(bd->shared_orphan_ring,
222 bucket_dequeue_buckets(struct bucket_data *bd, void **obj_table,
223 unsigned int n_buckets)
225 struct bucket_stack *cur_stack = bd->buckets[rte_lcore_id()];
226 unsigned int n_buckets_from_stack = RTE_MIN(n_buckets, cur_stack->top);
227 void **obj_table_base = obj_table;
229 n_buckets -= n_buckets_from_stack;
230 while (n_buckets_from_stack-- > 0) {
231 void *obj = bucket_stack_pop_unsafe(cur_stack);
233 obj_table = bucket_fill_obj_table(bd, &obj, obj_table,
236 while (n_buckets-- > 0) {
237 struct bucket_header *hdr;
239 if (unlikely(rte_ring_dequeue(bd->shared_bucket_ring,
240 (void **)&hdr) != 0)) {
242 * Return the already-dequeued buffers
243 * back to the mempool
245 bucket_enqueue(bd->pool, obj_table_base,
246 obj_table - obj_table_base);
250 hdr->lcore_id = rte_lcore_id();
251 obj_table = bucket_fill_obj_table(bd, (void **)&hdr,
260 bucket_adopt_orphans(struct bucket_data *bd)
263 struct rte_ring *adopt_ring =
264 bd->adoption_buffer_rings[rte_lcore_id()];
266 if (unlikely(!rte_ring_empty(adopt_ring))) {
269 while (rte_ring_sc_dequeue(adopt_ring, &orphan) == 0) {
270 rc = bucket_enqueue_single(bd, orphan);
278 bucket_dequeue(struct rte_mempool *mp, void **obj_table, unsigned int n)
280 struct bucket_data *bd = mp->pool_data;
281 unsigned int n_buckets = n / bd->obj_per_bucket;
282 unsigned int n_orphans = n - n_buckets * bd->obj_per_bucket;
285 bucket_adopt_orphans(bd);
287 if (unlikely(n_orphans > 0)) {
288 rc = bucket_dequeue_orphans(bd, obj_table +
289 (n_buckets * bd->obj_per_bucket),
295 if (likely(n_buckets > 0)) {
296 rc = bucket_dequeue_buckets(bd, obj_table, n_buckets);
297 if (unlikely(rc != 0) && n_orphans > 0) {
298 rte_ring_enqueue_bulk(bd->shared_orphan_ring,
299 obj_table + (n_buckets *
309 bucket_dequeue_contig_blocks(struct rte_mempool *mp, void **first_obj_table,
312 struct bucket_data *bd = mp->pool_data;
313 const uint32_t header_size = bd->header_size;
314 struct bucket_stack *cur_stack = bd->buckets[rte_lcore_id()];
315 unsigned int n_buckets_from_stack = RTE_MIN(n, cur_stack->top);
316 struct bucket_header *hdr;
317 void **first_objp = first_obj_table;
319 bucket_adopt_orphans(bd);
321 n -= n_buckets_from_stack;
322 while (n_buckets_from_stack-- > 0) {
323 hdr = bucket_stack_pop_unsafe(cur_stack);
324 *first_objp++ = (uint8_t *)hdr + header_size;
327 if (unlikely(rte_ring_dequeue_bulk(bd->shared_bucket_ring,
328 first_objp, n, NULL) != n)) {
329 /* Return the already dequeued buckets */
330 while (first_objp-- != first_obj_table) {
331 bucket_stack_push(cur_stack,
332 (uint8_t *)*first_objp -
339 hdr = (struct bucket_header *)*first_objp;
340 hdr->lcore_id = rte_lcore_id();
341 *first_objp++ = (uint8_t *)hdr + header_size;
349 count_underfilled_buckets(struct rte_mempool *mp,
351 struct rte_mempool_memhdr *memhdr,
352 __rte_unused unsigned int mem_idx)
354 unsigned int *pcount = opaque;
355 const struct bucket_data *bd = mp->pool_data;
356 unsigned int bucket_page_sz =
357 (unsigned int)(~bd->bucket_page_mask + 1);
361 align = (uintptr_t)RTE_PTR_ALIGN_CEIL(memhdr->addr, bucket_page_sz) -
362 (uintptr_t)memhdr->addr;
364 for (iter = (uint8_t *)memhdr->addr + align;
365 iter < (uint8_t *)memhdr->addr + memhdr->len;
366 iter += bucket_page_sz) {
367 struct bucket_header *hdr = (struct bucket_header *)iter;
369 *pcount += hdr->fill_cnt;
374 bucket_get_count(const struct rte_mempool *mp)
376 const struct bucket_data *bd = mp->pool_data;
378 bd->obj_per_bucket * rte_ring_count(bd->shared_bucket_ring) +
379 rte_ring_count(bd->shared_orphan_ring);
382 for (i = 0; i < RTE_MAX_LCORE; i++) {
383 if (!rte_lcore_is_enabled(i))
385 count += bd->obj_per_bucket * bd->buckets[i]->top +
386 rte_ring_count(bd->adoption_buffer_rings[i]);
389 rte_mempool_mem_iter((struct rte_mempool *)(uintptr_t)mp,
390 count_underfilled_buckets, &count);
396 bucket_alloc(struct rte_mempool *mp)
400 char rg_name[RTE_RING_NAMESIZE];
401 struct bucket_data *bd;
403 unsigned int bucket_header_size;
405 bd = rte_zmalloc_socket("bucket_pool", sizeof(*bd),
406 RTE_CACHE_LINE_SIZE, mp->socket_id);
409 goto no_mem_for_data;
412 if (mp->flags & MEMPOOL_F_NO_CACHE_ALIGN)
413 bucket_header_size = sizeof(struct bucket_header);
415 bucket_header_size = RTE_CACHE_LINE_SIZE;
416 RTE_BUILD_BUG_ON(sizeof(struct bucket_header) > RTE_CACHE_LINE_SIZE);
417 bd->header_size = mp->header_size + bucket_header_size;
418 bd->total_elt_size = mp->header_size + mp->elt_size + mp->trailer_size;
419 bd->bucket_mem_size = RTE_DRIVER_MEMPOOL_BUCKET_SIZE_KB * 1024;
420 bd->obj_per_bucket = (bd->bucket_mem_size - bucket_header_size) /
422 bd->bucket_page_mask = ~(rte_align64pow2(bd->bucket_mem_size) - 1);
423 /* eventually this should be a tunable parameter */
424 bd->bucket_stack_thresh = (mp->size / bd->obj_per_bucket) * 4 / 3;
426 if (mp->flags & MEMPOOL_F_SP_PUT)
427 rg_flags |= RING_F_SP_ENQ;
428 if (mp->flags & MEMPOOL_F_SC_GET)
429 rg_flags |= RING_F_SC_DEQ;
431 for (i = 0; i < RTE_MAX_LCORE; i++) {
432 if (!rte_lcore_is_enabled(i))
435 bucket_stack_create(mp, mp->size / bd->obj_per_bucket);
436 if (bd->buckets[i] == NULL) {
438 goto no_mem_for_stacks;
440 rc = snprintf(rg_name, sizeof(rg_name),
441 RTE_MEMPOOL_MZ_FORMAT ".a%u", mp->name, i);
442 if (rc < 0 || rc >= (int)sizeof(rg_name)) {
444 goto no_mem_for_stacks;
446 bd->adoption_buffer_rings[i] =
447 rte_ring_create(rg_name, rte_align32pow2(mp->size + 1),
449 rg_flags | RING_F_SC_DEQ);
450 if (bd->adoption_buffer_rings[i] == NULL) {
452 goto no_mem_for_stacks;
456 rc = snprintf(rg_name, sizeof(rg_name),
457 RTE_MEMPOOL_MZ_FORMAT ".0", mp->name);
458 if (rc < 0 || rc >= (int)sizeof(rg_name)) {
460 goto invalid_shared_orphan_ring;
462 bd->shared_orphan_ring =
463 rte_ring_create(rg_name, rte_align32pow2(mp->size + 1),
464 mp->socket_id, rg_flags);
465 if (bd->shared_orphan_ring == NULL) {
467 goto cannot_create_shared_orphan_ring;
470 rc = snprintf(rg_name, sizeof(rg_name),
471 RTE_MEMPOOL_MZ_FORMAT ".1", mp->name);
472 if (rc < 0 || rc >= (int)sizeof(rg_name)) {
474 goto invalid_shared_bucket_ring;
476 bd->shared_bucket_ring =
477 rte_ring_create(rg_name,
478 rte_align32pow2((mp->size + 1) /
480 mp->socket_id, rg_flags);
481 if (bd->shared_bucket_ring == NULL) {
483 goto cannot_create_shared_bucket_ring;
490 cannot_create_shared_bucket_ring:
491 invalid_shared_bucket_ring:
492 rte_ring_free(bd->shared_orphan_ring);
493 cannot_create_shared_orphan_ring:
494 invalid_shared_orphan_ring:
496 for (i = 0; i < RTE_MAX_LCORE; i++) {
497 rte_free(bd->buckets[i]);
498 rte_ring_free(bd->adoption_buffer_rings[i]);
507 bucket_free(struct rte_mempool *mp)
510 struct bucket_data *bd = mp->pool_data;
515 for (i = 0; i < RTE_MAX_LCORE; i++) {
516 rte_free(bd->buckets[i]);
517 rte_ring_free(bd->adoption_buffer_rings[i]);
520 rte_ring_free(bd->shared_orphan_ring);
521 rte_ring_free(bd->shared_bucket_ring);
527 bucket_calc_mem_size(const struct rte_mempool *mp, uint32_t obj_num,
528 __rte_unused uint32_t pg_shift, size_t *min_total_elt_size,
531 struct bucket_data *bd = mp->pool_data;
532 unsigned int bucket_page_sz;
537 bucket_page_sz = rte_align32pow2(bd->bucket_mem_size);
538 *align = bucket_page_sz;
539 *min_total_elt_size = bucket_page_sz;
541 * Each bucket occupies its own block aligned to
542 * bucket_page_sz, so the required amount of memory is
543 * a multiple of bucket_page_sz.
544 * We also need extra space for a bucket header
546 return ((obj_num + bd->obj_per_bucket - 1) /
547 bd->obj_per_bucket) * bucket_page_sz;
551 bucket_populate(struct rte_mempool *mp, unsigned int max_objs,
552 void *vaddr, rte_iova_t iova, size_t len,
553 rte_mempool_populate_obj_cb_t *obj_cb, void *obj_cb_arg)
555 struct bucket_data *bd = mp->pool_data;
556 unsigned int bucket_page_sz;
557 unsigned int bucket_header_sz;
566 bucket_page_sz = rte_align32pow2(bd->bucket_mem_size);
567 align = RTE_PTR_ALIGN_CEIL((uintptr_t)vaddr, bucket_page_sz) -
570 bucket_header_sz = bd->header_size - mp->header_size;
571 if (iova != RTE_BAD_IOVA)
572 iova += align + bucket_header_sz;
574 for (iter = (uint8_t *)vaddr + align, n_objs = 0;
575 iter < (uint8_t *)vaddr + len && n_objs < max_objs;
576 iter += bucket_page_sz) {
577 struct bucket_header *hdr = (struct bucket_header *)iter;
578 unsigned int chunk_len = bd->bucket_mem_size;
580 if ((size_t)(iter - (uint8_t *)vaddr) + chunk_len > len)
581 chunk_len = len - (iter - (uint8_t *)vaddr);
582 if (chunk_len <= bucket_header_sz)
584 chunk_len -= bucket_header_sz;
587 hdr->lcore_id = LCORE_ID_ANY;
588 rc = rte_mempool_op_populate_default(mp,
589 RTE_MIN(bd->obj_per_bucket,
591 iter + bucket_header_sz,
597 if (iova != RTE_BAD_IOVA)
598 iova += bucket_page_sz;
605 bucket_get_info(const struct rte_mempool *mp, struct rte_mempool_info *info)
607 struct bucket_data *bd = mp->pool_data;
609 info->contig_block_size = bd->obj_per_bucket;
614 static const struct rte_mempool_ops ops_bucket = {
616 .alloc = bucket_alloc,
618 .enqueue = bucket_enqueue,
619 .dequeue = bucket_dequeue,
620 .get_count = bucket_get_count,
621 .calc_mem_size = bucket_calc_mem_size,
622 .populate = bucket_populate,
623 .get_info = bucket_get_info,
624 .dequeue_contig_blocks = bucket_dequeue_contig_blocks,
628 MEMPOOL_REGISTER_OPS(ops_bucket);