numam-dpdk/lib/librte_mempool/rte_mempool.c

/*-
 *   BSD LICENSE
 *
 *   Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
 *   Copyright(c) 2016 6WIND S.A.
 *   All rights reserved.
 *
 *   Redistribution and use in source and binary forms, with or without
 *   modification, are permitted provided that the following conditions
 *   are met:
 *
 *     * Redistributions of source code must retain the above copyright
 *       notice, this list of conditions and the following disclaimer.
 *     * Redistributions in binary form must reproduce the above copyright
 *       notice, this list of conditions and the following disclaimer in
 *       the documentation and/or other materials provided with the
 *       distribution.
 *     * Neither the name of Intel Corporation nor the names of its
 *       contributors may be used to endorse or promote products derived
 *       from this software without specific prior written permission.
 *
 *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
 *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
 *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
 *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

#include <stdio.h>
#include <string.h>
#include <stdint.h>
#include <stdarg.h>
#include <unistd.h>
#include <inttypes.h>
#include <errno.h>
#include <sys/queue.h>

#include <rte_common.h>
#include <rte_log.h>
#include <rte_debug.h>
#include <rte_memory.h>
#include <rte_memzone.h>
#include <rte_malloc.h>
#include <rte_atomic.h>
#include <rte_launch.h>
#include <rte_eal.h>
#include <rte_eal_memconfig.h>
#include <rte_per_lcore.h>
#include <rte_lcore.h>
#include <rte_branch_prediction.h>
#include <rte_ring.h>
#include <rte_errno.h>
#include <rte_string_fns.h>
#include <rte_spinlock.h>

#include "rte_mempool.h"

TAILQ_HEAD(rte_mempool_list, rte_tailq_entry);

static struct rte_tailq_elem rte_mempool_tailq = {
	.name = "RTE_MEMPOOL",
};
EAL_REGISTER_TAILQ(rte_mempool_tailq)

#define CACHE_FLUSHTHRESH_MULTIPLIER 1.5
#define CALC_CACHE_FLUSHTHRESH(c)	\
	((typeof(c))((c) * CACHE_FLUSHTHRESH_MULTIPLIER))

/*
 * return the greatest common divisor between a and b (fast algorithm)
 *
 */
static unsigned get_gcd(unsigned a, unsigned b)
{
	unsigned c;

	if (0 == a)
		return b;
	if (0 == b)
		return a;

	if (a < b) {
		c = a;
		a = b;
		b = c;
	}

	while (b != 0) {
		c = a % b;
		a = b;
		b = c;
	}

	return a;
}

/*
 * Depending on memory configuration, objects addresses are spread
 * between channels and ranks in RAM: the pool allocator will add
 * padding between objects. This function return the new size of the
 * object.
 */
static unsigned optimize_object_size(unsigned obj_size)
{
	unsigned nrank, nchan;
	unsigned new_obj_size;

	/* get number of channels */
	nchan = rte_memory_get_nchannel();
	if (nchan == 0)
		nchan = 4;

	nrank = rte_memory_get_nrank();
	if (nrank == 0)
		nrank = 1;

	/* process new object size */
	new_obj_size = (obj_size + RTE_MEMPOOL_ALIGN_MASK) / RTE_MEMPOOL_ALIGN;
	while (get_gcd(new_obj_size, nrank * nchan) != 1)
		new_obj_size++;
	return new_obj_size * RTE_MEMPOOL_ALIGN;
}

static void
mempool_add_elem(struct rte_mempool *mp, void *obj, phys_addr_t physaddr)
{
	struct rte_mempool_objhdr *hdr;
	struct rte_mempool_objtlr *tlr __rte_unused;

	/* set mempool ptr in header */
	hdr = RTE_PTR_SUB(obj, sizeof(*hdr));
	hdr->mp = mp;
	hdr->physaddr = physaddr;
	STAILQ_INSERT_TAIL(&mp->elt_list, hdr, next);
	mp->populated_size++;

#ifdef RTE_LIBRTE_MEMPOOL_DEBUG
	hdr->cookie = RTE_MEMPOOL_HEADER_COOKIE2;
	tlr = __mempool_get_trailer(obj);
	tlr->cookie = RTE_MEMPOOL_TRAILER_COOKIE;
#endif

	/* enqueue in ring */
	rte_ring_sp_enqueue(mp->ring, obj);
}

/* call obj_cb() for each mempool element */
uint32_t
rte_mempool_obj_iter(struct rte_mempool *mp,
	rte_mempool_obj_cb_t *obj_cb, void *obj_cb_arg)
{
	struct rte_mempool_objhdr *hdr;
	void *obj;
	unsigned n = 0;

	STAILQ_FOREACH(hdr, &mp->elt_list, next) {
		obj = (char *)hdr + sizeof(*hdr);
		obj_cb(mp, obj_cb_arg, obj, n);
		n++;
	}

	return n;
}

/* call mem_cb() for each mempool memory chunk */
uint32_t
rte_mempool_mem_iter(struct rte_mempool *mp,
	rte_mempool_mem_cb_t *mem_cb, void *mem_cb_arg)
{
	struct rte_mempool_memhdr *hdr;
	unsigned n = 0;

	STAILQ_FOREACH(hdr, &mp->mem_list, next) {
		mem_cb(mp, mem_cb_arg, hdr, n);
		n++;
	}

	return n;
}

/* get the header, trailer and total size of a mempool element. */
uint32_t
rte_mempool_calc_obj_size(uint32_t elt_size, uint32_t flags,
	struct rte_mempool_objsz *sz)
{
	struct rte_mempool_objsz lsz;

	sz = (sz != NULL) ? sz : &lsz;

	sz->header_size = sizeof(struct rte_mempool_objhdr);
	if ((flags & MEMPOOL_F_NO_CACHE_ALIGN) == 0)
		sz->header_size = RTE_ALIGN_CEIL(sz->header_size,
			RTE_MEMPOOL_ALIGN);

	sz->trailer_size = sizeof(struct rte_mempool_objtlr);

	/* element size is 8 bytes-aligned at least */
	sz->elt_size = RTE_ALIGN_CEIL(elt_size, sizeof(uint64_t));

	/* expand trailer to next cache line */
	if ((flags & MEMPOOL_F_NO_CACHE_ALIGN) == 0) {
		sz->total_size = sz->header_size + sz->elt_size +
			sz->trailer_size;
		sz->trailer_size += ((RTE_MEMPOOL_ALIGN -
				  (sz->total_size & RTE_MEMPOOL_ALIGN_MASK)) &
				 RTE_MEMPOOL_ALIGN_MASK);
	}

	/*
	 * increase trailer to add padding between objects in order to
	 * spread them across memory channels/ranks
	 */
	if ((flags & MEMPOOL_F_NO_SPREAD) == 0) {
		unsigned new_size;
		new_size = optimize_object_size(sz->header_size + sz->elt_size +
			sz->trailer_size);
		sz->trailer_size = new_size - sz->header_size - sz->elt_size;
	}

	if (! rte_eal_has_hugepages()) {
		/*
		 * compute trailer size so that pool elements fit exactly in
		 * a standard page
		 */
		int page_size = getpagesize();
		int new_size = page_size - sz->header_size - sz->elt_size;
		if (new_size < 0 || (unsigned int)new_size < sz->trailer_size) {
			printf("When hugepages are disabled, pool objects "
			       "can't exceed PAGE_SIZE: %d + %d + %d > %d\n",
			       sz->header_size, sz->elt_size, sz->trailer_size,
			       page_size);
			return 0;
		}
		sz->trailer_size = new_size;
	}

	/* this is the size of an object, including header and trailer */
	sz->total_size = sz->header_size + sz->elt_size + sz->trailer_size;

	return sz->total_size;
}


/*
 * Calculate maximum amount of memory required to store given number of objects.
 */
size_t
rte_mempool_xmem_size(uint32_t elt_num, size_t total_elt_sz, uint32_t pg_shift)
{
	size_t n, pg_num, pg_sz, sz;

	pg_sz = (size_t)1 << pg_shift;

	if ((n = pg_sz / total_elt_sz) > 0) {
		pg_num = (elt_num + n - 1) / n;
		sz = pg_num << pg_shift;
	} else {
		sz = RTE_ALIGN_CEIL(total_elt_sz, pg_sz) * elt_num;
	}

	return sz;
}

/*
 * Calculate how much memory would be actually required with the
 * given memory footprint to store required number of elements.
 */
ssize_t
rte_mempool_xmem_usage(__rte_unused void *vaddr, uint32_t elt_num,
	size_t total_elt_sz, const phys_addr_t paddr[], uint32_t pg_num,
	uint32_t pg_shift)
{
	uint32_t elt_cnt = 0;
	phys_addr_t start, end;
	uint32_t paddr_idx;
	size_t pg_sz = (size_t)1 << pg_shift;

	/* if paddr is NULL, assume contiguous memory */
	if (paddr == NULL) {
		start = 0;
		end = pg_sz * pg_num;
		paddr_idx = pg_num;
	} else {
		start = paddr[0];
		end = paddr[0] + pg_sz;
		paddr_idx = 1;
	}
	while (elt_cnt < elt_num) {

		if (end - start >= total_elt_sz) {
			/* enough contiguous memory, add an object */
			start += total_elt_sz;
			elt_cnt++;
		} else if (paddr_idx < pg_num) {
			/* no room to store one obj, add a page */
			if (end == paddr[paddr_idx]) {
				end += pg_sz;
			} else {
				start = paddr[paddr_idx];
				end = paddr[paddr_idx] + pg_sz;
			}
			paddr_idx++;

		} else {
			/* no more page, return how many elements fit */
			return -(size_t)elt_cnt;
		}
	}

	return (size_t)paddr_idx << pg_shift;
}

#ifndef RTE_LIBRTE_XEN_DOM0
/* stub if DOM0 support not configured */
struct rte_mempool *
rte_dom0_mempool_create(const char *name __rte_unused,
			unsigned n __rte_unused,
			unsigned elt_size __rte_unused,
			unsigned cache_size __rte_unused,
			unsigned private_data_size __rte_unused,
			rte_mempool_ctor_t *mp_init __rte_unused,
			void *mp_init_arg __rte_unused,
			rte_mempool_obj_ctor_t *obj_init __rte_unused,
			void *obj_init_arg __rte_unused,
			int socket_id __rte_unused,
			unsigned flags __rte_unused)
{
	rte_errno = EINVAL;
	return NULL;
}
#endif

/* create the mempool */
struct rte_mempool *
rte_mempool_create(const char *name, unsigned n, unsigned elt_size,
		   unsigned cache_size, unsigned private_data_size,
		   rte_mempool_ctor_t *mp_init, void *mp_init_arg,
		   rte_mempool_obj_ctor_t *obj_init, void *obj_init_arg,
		   int socket_id, unsigned flags)
{
	if (rte_xen_dom0_supported())
		return rte_dom0_mempool_create(name, n, elt_size,
					       cache_size, private_data_size,
					       mp_init, mp_init_arg,
					       obj_init, obj_init_arg,
					       socket_id, flags);
	else
		return rte_mempool_xmem_create(name, n, elt_size,
					       cache_size, private_data_size,
					       mp_init, mp_init_arg,
					       obj_init, obj_init_arg,
					       socket_id, flags,
					       NULL, NULL, MEMPOOL_PG_NUM_DEFAULT,
					       MEMPOOL_PG_SHIFT_MAX);
}

/* create the internal ring */
static int
rte_mempool_ring_create(struct rte_mempool *mp)
{
	int rg_flags = 0;
	char rg_name[RTE_RING_NAMESIZE];
	struct rte_ring *r;

	snprintf(rg_name, sizeof(rg_name), RTE_MEMPOOL_MZ_FORMAT, mp->name);

	/* ring flags */
	if (mp->flags & MEMPOOL_F_SP_PUT)
		rg_flags |= RING_F_SP_ENQ;
	if (mp->flags & MEMPOOL_F_SC_GET)
		rg_flags |= RING_F_SC_DEQ;

	/* Allocate the ring that will be used to store objects.
	 * Ring functions will return appropriate errors if we are
	 * running as a secondary process etc., so no checks made
	 * in this function for that condition.
	 */
	r = rte_ring_create(rg_name, rte_align32pow2(mp->size + 1),
		mp->socket_id, rg_flags);
	if (r == NULL)
		return -rte_errno;

	mp->ring = r;
	return 0;
}

/* Free memory chunks used by a mempool. Objects must be in pool */
static void
rte_mempool_free_memchunks(struct rte_mempool *mp)
{
	struct rte_mempool_memhdr *memhdr;
	void *elt;

	while (!STAILQ_EMPTY(&mp->elt_list)) {
		rte_ring_sc_dequeue(mp->ring, &elt);
		(void)elt;
		STAILQ_REMOVE_HEAD(&mp->elt_list, next);
		mp->populated_size--;
	}

	while (!STAILQ_EMPTY(&mp->mem_list)) {
		memhdr = STAILQ_FIRST(&mp->mem_list);
		STAILQ_REMOVE_HEAD(&mp->mem_list, next);
		rte_free(memhdr);
		mp->nb_mem_chunks--;
	}
}

/* Add objects in the pool, using a physically contiguous memory
 * zone. Return the number of objects added, or a negative value
 * on error.
 */
static int
rte_mempool_populate_phys(struct rte_mempool *mp, char *vaddr,
	phys_addr_t paddr, size_t len)
{
	unsigned total_elt_sz;
	unsigned i = 0;
	size_t off;
	struct rte_mempool_memhdr *memhdr;

	/* mempool is already populated */
	if (mp->populated_size >= mp->size)
		return -ENOSPC;

	total_elt_sz = mp->header_size + mp->elt_size + mp->trailer_size;

	memhdr = rte_zmalloc("MEMPOOL_MEMHDR", sizeof(*memhdr), 0);
	if (memhdr == NULL)
		return -ENOMEM;

	memhdr->mp = mp;
	memhdr->addr = vaddr;
	memhdr->phys_addr = paddr;
	memhdr->len = len;

	if (mp->flags & MEMPOOL_F_NO_CACHE_ALIGN)
		off = RTE_PTR_ALIGN_CEIL(vaddr, 8) - vaddr;
	else
		off = RTE_PTR_ALIGN_CEIL(vaddr, RTE_CACHE_LINE_SIZE) - vaddr;

	while (off + total_elt_sz <= len && mp->populated_size < mp->size) {
		off += mp->header_size;
		mempool_add_elem(mp, (char *)vaddr + off, paddr + off);
		off += mp->elt_size + mp->trailer_size;
		i++;
	}

	/* not enough room to store one object */
	if (i == 0)
		return -EINVAL;

	STAILQ_INSERT_TAIL(&mp->mem_list, memhdr, next);
	mp->nb_mem_chunks++;
	return i;
}

/* Add objects in the pool, using a table of physical pages. Return the
 * number of objects added, or a negative value on error.
 */
static int
rte_mempool_populate_phys_tab(struct rte_mempool *mp, char *vaddr,
	const phys_addr_t paddr[], uint32_t pg_num, uint32_t pg_shift)
{
	uint32_t i, n;
	int ret, cnt = 0;
	size_t pg_sz = (size_t)1 << pg_shift;

	/* mempool must not be populated */
	if (mp->nb_mem_chunks != 0)
		return -EEXIST;

	for (i = 0; i < pg_num && mp->populated_size < mp->size; i += n) {

		/* populate with the largest group of contiguous pages */
		for (n = 1; (i + n) < pg_num &&
			     paddr[i] + pg_sz == paddr[i+n]; n++)
			;

		ret = rte_mempool_populate_phys(mp, vaddr + i * pg_sz,
			paddr[i], n * pg_sz);
		if (ret < 0) {
			rte_mempool_free_memchunks(mp);
			return ret;
		}
		cnt += ret;
	}
	return cnt;
}

/*
 * Create the mempool over already allocated chunk of memory.
 * That external memory buffer can consists of physically disjoint pages.
 * Setting vaddr to NULL, makes mempool to fallback to original behaviour
 * and allocate space for mempool and it's elements as one big chunk of
 * physically continuos memory.
 * */
struct rte_mempool *
rte_mempool_xmem_create(const char *name, unsigned n, unsigned elt_size,
		unsigned cache_size, unsigned private_data_size,
		rte_mempool_ctor_t *mp_init, void *mp_init_arg,
		rte_mempool_obj_cb_t *obj_init, void *obj_init_arg,
		int socket_id, unsigned flags, void *vaddr,
		const phys_addr_t paddr[], uint32_t pg_num, uint32_t pg_shift)
{
	char mz_name[RTE_MEMZONE_NAMESIZE];
	struct rte_mempool_list *mempool_list;
	struct rte_mempool *mp = NULL;
	struct rte_tailq_entry *te = NULL;
	const struct rte_memzone *mz;
	size_t mempool_size;
	int mz_flags = RTE_MEMZONE_1GB|RTE_MEMZONE_SIZE_HINT_ONLY;
	void *obj;
	struct rte_mempool_objsz objsz;
	void *startaddr;
	int page_size = getpagesize();
	int ret;

	/* compilation-time checks */
	RTE_BUILD_BUG_ON((sizeof(struct rte_mempool) &
			  RTE_CACHE_LINE_MASK) != 0);
	RTE_BUILD_BUG_ON((sizeof(struct rte_mempool_cache) &
			  RTE_CACHE_LINE_MASK) != 0);
#ifdef RTE_LIBRTE_MEMPOOL_DEBUG
	RTE_BUILD_BUG_ON((sizeof(struct rte_mempool_debug_stats) &
			  RTE_CACHE_LINE_MASK) != 0);
	RTE_BUILD_BUG_ON((offsetof(struct rte_mempool, stats) &
			  RTE_CACHE_LINE_MASK) != 0);
#endif

	mempool_list = RTE_TAILQ_CAST(rte_mempool_tailq.head, rte_mempool_list);

	/* asked cache too big */
	if (cache_size > RTE_MEMPOOL_CACHE_MAX_SIZE ||
	    CALC_CACHE_FLUSHTHRESH(cache_size) > n) {
		rte_errno = EINVAL;
		return NULL;
	}

	/* check that we have both VA and PA */
	if (vaddr != NULL && paddr == NULL) {
		rte_errno = EINVAL;
		return NULL;
	}

	/* Check that pg_num and pg_shift parameters are valid. */
	if (pg_num == 0 || pg_shift > MEMPOOL_PG_SHIFT_MAX) {
		rte_errno = EINVAL;
		return NULL;
	}

	/* "no cache align" imply "no spread" */
	if (flags & MEMPOOL_F_NO_CACHE_ALIGN)
		flags |= MEMPOOL_F_NO_SPREAD;

	/* calculate mempool object sizes. */
	if (!rte_mempool_calc_obj_size(elt_size, flags, &objsz)) {
		rte_errno = EINVAL;
		return NULL;
	}

	rte_rwlock_write_lock(RTE_EAL_MEMPOOL_RWLOCK);

	/*
	 * reserve a memory zone for this mempool: private data is
	 * cache-aligned
	 */
	private_data_size = (private_data_size +
			     RTE_MEMPOOL_ALIGN_MASK) & (~RTE_MEMPOOL_ALIGN_MASK);

	if (! rte_eal_has_hugepages()) {
		/*
		 * expand private data size to a whole page, so that the
		 * first pool element will start on a new standard page
		 */
		int head = sizeof(struct rte_mempool);
		int new_size = (private_data_size + head) % page_size;
		if (new_size)
			private_data_size += page_size - new_size;
	}

	/* try to allocate tailq entry */
	te = rte_zmalloc("MEMPOOL_TAILQ_ENTRY", sizeof(*te), 0);
	if (te == NULL) {
		RTE_LOG(ERR, MEMPOOL, "Cannot allocate tailq entry!\n");
		goto exit_unlock;
	}

	/*
	 * If user provided an external memory buffer, then use it to
	 * store mempool objects. Otherwise reserve a memzone that is large
	 * enough to hold mempool header and metadata plus mempool objects.
	 */
	mempool_size = MEMPOOL_HEADER_SIZE(mp, cache_size);
	mempool_size += private_data_size;
	mempool_size = RTE_ALIGN_CEIL(mempool_size, RTE_MEMPOOL_ALIGN);
	if (vaddr == NULL)
		mempool_size += (size_t)objsz.total_size * n;

	if (! rte_eal_has_hugepages()) {
		/*
		 * we want the memory pool to start on a page boundary,
		 * because pool elements crossing page boundaries would
		 * result in discontiguous physical addresses
		 */
		mempool_size += page_size;
	}

	snprintf(mz_name, sizeof(mz_name), RTE_MEMPOOL_MZ_FORMAT, name);

	mz = rte_memzone_reserve(mz_name, mempool_size, socket_id, mz_flags);
	if (mz == NULL)
		goto exit_unlock;

	if (rte_eal_has_hugepages()) {
		startaddr = (void*)mz->addr;
	} else {
		/* align memory pool start address on a page boundary */
		unsigned long addr = (unsigned long)mz->addr;
		if (addr & (page_size - 1)) {
			addr += page_size;
			addr &= ~(page_size - 1);
		}
		startaddr = (void*)addr;
	}

	/* init the mempool structure */
	mp = startaddr;
	memset(mp, 0, sizeof(*mp));
	snprintf(mp->name, sizeof(mp->name), "%s", name);
	mp->phys_addr = mz->phys_addr;
	mp->socket_id = socket_id;
	mp->size = n;
	mp->flags = flags;
	mp->elt_size = objsz.elt_size;
	mp->header_size = objsz.header_size;
	mp->trailer_size = objsz.trailer_size;
	mp->cache_size = cache_size;
	mp->cache_flushthresh = CALC_CACHE_FLUSHTHRESH(cache_size);
	mp->private_data_size = private_data_size;
	STAILQ_INIT(&mp->elt_list);
	STAILQ_INIT(&mp->mem_list);

	if (rte_mempool_ring_create(mp) < 0)
		goto exit_unlock;

	/*
	 * local_cache pointer is set even if cache_size is zero.
	 * The local_cache points to just past the elt_pa[] array.
	 */
	mp->local_cache = (struct rte_mempool_cache *)
		RTE_PTR_ADD(mp, MEMPOOL_HEADER_SIZE(mp, 0));

	/* call the initializer */
	if (mp_init)
		mp_init(mp, mp_init_arg);

	/* mempool elements allocated together with mempool */
	if (vaddr == NULL) {
		/* calculate address of the first elt for continuous mempool. */
		obj = (char *)mp + MEMPOOL_HEADER_SIZE(mp, cache_size) +
			private_data_size;
		obj = RTE_PTR_ALIGN_CEIL(obj, RTE_MEMPOOL_ALIGN);

		ret = rte_mempool_populate_phys(mp, obj,
			mp->phys_addr + ((char *)obj - (char *)mp),
			objsz.total_size * n);
		if (ret != (int)mp->size)
			goto exit_unlock;
	} else {
		ret = rte_mempool_populate_phys_tab(mp, vaddr,
			paddr, pg_num, pg_shift);
		if (ret != (int)mp->size)
			goto exit_unlock;
	}

	/* call the initializer */
	if (obj_init)
		rte_mempool_obj_iter(mp, obj_init, obj_init_arg);

	te->data = (void *) mp;

	rte_rwlock_write_lock(RTE_EAL_TAILQ_RWLOCK);
	TAILQ_INSERT_TAIL(mempool_list, te, next);
	rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);
	rte_rwlock_write_unlock(RTE_EAL_MEMPOOL_RWLOCK);

	return mp;

exit_unlock:
	rte_rwlock_write_unlock(RTE_EAL_MEMPOOL_RWLOCK);
	if (mp != NULL) {
		rte_mempool_free_memchunks(mp);
		rte_ring_free(mp->ring);
	}
	rte_free(te);

	return NULL;
}

/* Return the number of entries in the mempool */
unsigned
rte_mempool_count(const struct rte_mempool *mp)
{
	unsigned count;
	unsigned lcore_id;

	count = rte_ring_count(mp->ring);

	if (mp->cache_size == 0)
		return count;

	for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++)
		count += mp->local_cache[lcore_id].len;

	/*
	 * due to race condition (access to len is not locked), the
	 * total can be greater than size... so fix the result
	 */
	if (count > mp->size)
		return mp->size;
	return count;
}

/* dump the cache status */
static unsigned
rte_mempool_dump_cache(FILE *f, const struct rte_mempool *mp)
{
	unsigned lcore_id;
	unsigned count = 0;
	unsigned cache_count;

	fprintf(f, "  cache infos:\n");
	fprintf(f, "    cache_size=%"PRIu32"\n", mp->cache_size);

	if (mp->cache_size == 0)
		return count;

	for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
		cache_count = mp->local_cache[lcore_id].len;
		fprintf(f, "    cache_count[%u]=%u\n", lcore_id, cache_count);
		count += cache_count;
	}
	fprintf(f, "    total_cache_count=%u\n", count);
	return count;
}

#ifndef __INTEL_COMPILER
#pragma GCC diagnostic ignored "-Wcast-qual"
#endif

/* check and update cookies or panic (internal) */
void rte_mempool_check_cookies(const struct rte_mempool *mp,
	void * const *obj_table_const, unsigned n, int free)
{
#ifdef RTE_LIBRTE_MEMPOOL_DEBUG
	struct rte_mempool_objhdr *hdr;
	struct rte_mempool_objtlr *tlr;
	uint64_t cookie;
	void *tmp;
	void *obj;
	void **obj_table;

	/* Force to drop the "const" attribute. This is done only when
	 * DEBUG is enabled */
	tmp = (void *) obj_table_const;
	obj_table = (void **) tmp;

	while (n--) {
		obj = obj_table[n];

		if (rte_mempool_from_obj(obj) != mp)
			rte_panic("MEMPOOL: object is owned by another "
				  "mempool\n");

		hdr = __mempool_get_header(obj);
		cookie = hdr->cookie;

		if (free == 0) {
			if (cookie != RTE_MEMPOOL_HEADER_COOKIE1) {
				rte_log_set_history(0);
				RTE_LOG(CRIT, MEMPOOL,
					"obj=%p, mempool=%p, cookie=%" PRIx64 "\n",
					obj, (const void *) mp, cookie);
				rte_panic("MEMPOOL: bad header cookie (put)\n");
			}
			hdr->cookie = RTE_MEMPOOL_HEADER_COOKIE2;
		} else if (free == 1) {
			if (cookie != RTE_MEMPOOL_HEADER_COOKIE2) {
				rte_log_set_history(0);
				RTE_LOG(CRIT, MEMPOOL,
					"obj=%p, mempool=%p, cookie=%" PRIx64 "\n",
					obj, (const void *) mp, cookie);
				rte_panic("MEMPOOL: bad header cookie (get)\n");
			}
			hdr->cookie = RTE_MEMPOOL_HEADER_COOKIE1;
		} else if (free == 2) {
			if (cookie != RTE_MEMPOOL_HEADER_COOKIE1 &&
			    cookie != RTE_MEMPOOL_HEADER_COOKIE2) {
				rte_log_set_history(0);
				RTE_LOG(CRIT, MEMPOOL,
					"obj=%p, mempool=%p, cookie=%" PRIx64 "\n",
					obj, (const void *) mp, cookie);
				rte_panic("MEMPOOL: bad header cookie (audit)\n");
			}
		}
		tlr = __mempool_get_trailer(obj);
		cookie = tlr->cookie;
		if (cookie != RTE_MEMPOOL_TRAILER_COOKIE) {
			rte_log_set_history(0);
			RTE_LOG(CRIT, MEMPOOL,
				"obj=%p, mempool=%p, cookie=%" PRIx64 "\n",
				obj, (const void *) mp, cookie);
			rte_panic("MEMPOOL: bad trailer cookie\n");
		}
	}
#else
	RTE_SET_USED(mp);
	RTE_SET_USED(obj_table_const);
	RTE_SET_USED(n);
	RTE_SET_USED(free);
#endif
}

#ifdef RTE_LIBRTE_MEMPOOL_DEBUG
static void
mempool_obj_audit(struct rte_mempool *mp, __rte_unused void *opaque,
	void *obj, __rte_unused unsigned idx)
{
	__mempool_check_cookies(mp, &obj, 1, 2);
}

static void
mempool_audit_cookies(struct rte_mempool *mp)
{
	unsigned num;

	num = rte_mempool_obj_iter(mp, mempool_obj_audit, NULL);
	if (num != mp->size) {
		rte_panic("rte_mempool_obj_iter(mempool=%p, size=%u) "
			"iterated only over %u elements\n",
			mp, mp->size, num);
	}
}
#else
#define mempool_audit_cookies(mp) do {} while(0)
#endif

#ifndef __INTEL_COMPILER
#pragma GCC diagnostic error "-Wcast-qual"
#endif

/* check cookies before and after objects */
static void
mempool_audit_cache(const struct rte_mempool *mp)
{
	/* check cache size consistency */
	unsigned lcore_id;

	if (mp->cache_size == 0)
		return;

	for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
		if (mp->local_cache[lcore_id].len > mp->cache_flushthresh) {
			RTE_LOG(CRIT, MEMPOOL, "badness on cache[%u]\n",
				lcore_id);
			rte_panic("MEMPOOL: invalid cache len\n");
		}
	}
}

/* check the consistency of mempool (size, cookies, ...) */
void
rte_mempool_audit(struct rte_mempool *mp)
{
	mempool_audit_cache(mp);
	mempool_audit_cookies(mp);

	/* For case where mempool DEBUG is not set, and cache size is 0 */
	RTE_SET_USED(mp);
}

/* dump the status of the mempool on the console */
void
rte_mempool_dump(FILE *f, struct rte_mempool *mp)
{
#ifdef RTE_LIBRTE_MEMPOOL_DEBUG
	struct rte_mempool_debug_stats sum;
	unsigned lcore_id;
#endif
	struct rte_mempool_memhdr *memhdr;
	unsigned common_count;
	unsigned cache_count;
	size_t mem_len = 0;

	RTE_ASSERT(f != NULL);
	RTE_ASSERT(mp != NULL);

	fprintf(f, "mempool <%s>@%p\n", mp->name, mp);
	fprintf(f, "  flags=%x\n", mp->flags);
	fprintf(f, "  ring=<%s>@%p\n", mp->ring->name, mp->ring);
	fprintf(f, "  phys_addr=0x%" PRIx64 "\n", mp->phys_addr);
	fprintf(f, "  nb_mem_chunks=%u\n", mp->nb_mem_chunks);
	fprintf(f, "  size=%"PRIu32"\n", mp->size);
	fprintf(f, "  populated_size=%"PRIu32"\n", mp->populated_size);
	fprintf(f, "  header_size=%"PRIu32"\n", mp->header_size);
	fprintf(f, "  elt_size=%"PRIu32"\n", mp->elt_size);
	fprintf(f, "  trailer_size=%"PRIu32"\n", mp->trailer_size);
	fprintf(f, "  total_obj_size=%"PRIu32"\n",
	       mp->header_size + mp->elt_size + mp->trailer_size);

	fprintf(f, "  private_data_size=%"PRIu32"\n", mp->private_data_size);

	STAILQ_FOREACH(memhdr, &mp->mem_list, next)
		mem_len += memhdr->len;
	if (mem_len != 0) {
		fprintf(f, "  avg bytes/object=%#Lf\n",
			(long double)mem_len / mp->size);
	}

	cache_count = rte_mempool_dump_cache(f, mp);
	common_count = rte_ring_count(mp->ring);
	if ((cache_count + common_count) > mp->size)
		common_count = mp->size - cache_count;
	fprintf(f, "  common_pool_count=%u\n", common_count);

	/* sum and dump statistics */
#ifdef RTE_LIBRTE_MEMPOOL_DEBUG
	memset(&sum, 0, sizeof(sum));
	for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
		sum.put_bulk += mp->stats[lcore_id].put_bulk;
		sum.put_objs += mp->stats[lcore_id].put_objs;
		sum.get_success_bulk += mp->stats[lcore_id].get_success_bulk;
		sum.get_success_objs += mp->stats[lcore_id].get_success_objs;
		sum.get_fail_bulk += mp->stats[lcore_id].get_fail_bulk;
		sum.get_fail_objs += mp->stats[lcore_id].get_fail_objs;
	}
	fprintf(f, "  stats:\n");
	fprintf(f, "    put_bulk=%"PRIu64"\n", sum.put_bulk);
	fprintf(f, "    put_objs=%"PRIu64"\n", sum.put_objs);
	fprintf(f, "    get_success_bulk=%"PRIu64"\n", sum.get_success_bulk);
	fprintf(f, "    get_success_objs=%"PRIu64"\n", sum.get_success_objs);
	fprintf(f, "    get_fail_bulk=%"PRIu64"\n", sum.get_fail_bulk);
	fprintf(f, "    get_fail_objs=%"PRIu64"\n", sum.get_fail_objs);
#else
	fprintf(f, "  no statistics available\n");
#endif

	rte_mempool_audit(mp);
}

/* dump the status of all mempools on the console */
void
rte_mempool_list_dump(FILE *f)
{
	struct rte_mempool *mp = NULL;
	struct rte_tailq_entry *te;
	struct rte_mempool_list *mempool_list;

	mempool_list = RTE_TAILQ_CAST(rte_mempool_tailq.head, rte_mempool_list);

	rte_rwlock_read_lock(RTE_EAL_MEMPOOL_RWLOCK);

	TAILQ_FOREACH(te, mempool_list, next) {
		mp = (struct rte_mempool *) te->data;
		rte_mempool_dump(f, mp);
	}

	rte_rwlock_read_unlock(RTE_EAL_MEMPOOL_RWLOCK);
}

/* search a mempool from its name */
struct rte_mempool *
rte_mempool_lookup(const char *name)
{
	struct rte_mempool *mp = NULL;
	struct rte_tailq_entry *te;
	struct rte_mempool_list *mempool_list;

	mempool_list = RTE_TAILQ_CAST(rte_mempool_tailq.head, rte_mempool_list);

	rte_rwlock_read_lock(RTE_EAL_MEMPOOL_RWLOCK);

	TAILQ_FOREACH(te, mempool_list, next) {
		mp = (struct rte_mempool *) te->data;
		if (strncmp(name, mp->name, RTE_MEMPOOL_NAMESIZE) == 0)
			break;
	}

	rte_rwlock_read_unlock(RTE_EAL_MEMPOOL_RWLOCK);

	if (te == NULL) {
		rte_errno = ENOENT;
		return NULL;
	}

	return mp;
}

void rte_mempool_walk(void (*func)(struct rte_mempool *, void *),
		      void *arg)
{
	struct rte_tailq_entry *te = NULL;
	struct rte_mempool_list *mempool_list;

	mempool_list = RTE_TAILQ_CAST(rte_mempool_tailq.head, rte_mempool_list);

	rte_rwlock_read_lock(RTE_EAL_MEMPOOL_RWLOCK);

	TAILQ_FOREACH(te, mempool_list, next) {
		(*func)((struct rte_mempool *) te->data, arg);
	}

	rte_rwlock_read_unlock(RTE_EAL_MEMPOOL_RWLOCK);
}