[ceph.git] / ceph / src / seastar / dpdk / lib / librte_ring / rte_ring.c

/* SPDX-License-Identifier: BSD-3-Clause
 *
 * Copyright (c) 2010-2015 Intel Corporation
 * Copyright (c) 2007,2008 Kip Macy kmacy@freebsd.org
 * All rights reserved.
 * Derived from FreeBSD's bufring.h
 * Used as BSD-3 Licensed with permission from Kip Macy.
 */

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

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

#include "rte_ring.h"

TAILQ_HEAD(rte_ring_list, rte_tailq_entry);

static struct rte_tailq_elem rte_ring_tailq = {
	.name = RTE_TAILQ_RING_NAME,
};
EAL_REGISTER_TAILQ(rte_ring_tailq)

/* true if x is a power of 2 */
#define POWEROF2(x) ((((x)-1) & (x)) == 0)

/* return the size of memory occupied by a ring */
ssize_t
rte_ring_get_memsize(unsigned count)
{
	ssize_t sz;

	/* count must be a power of 2 */
	if ((!POWEROF2(count)) || (count > RTE_RING_SZ_MASK )) {
		RTE_LOG(ERR, RING,
			"Requested size is invalid, must be power of 2, and "
			"do not exceed the size limit %u\n", RTE_RING_SZ_MASK);
		return -EINVAL;
	}

	sz = sizeof(struct rte_ring) + count * sizeof(void *);
	sz = RTE_ALIGN(sz, RTE_CACHE_LINE_SIZE);
	return sz;
}

int
rte_ring_init(struct rte_ring *r, const char *name, unsigned count,
	unsigned flags)
{
	int ret;

	/* compilation-time checks */
	RTE_BUILD_BUG_ON((sizeof(struct rte_ring) &
			  RTE_CACHE_LINE_MASK) != 0);
	RTE_BUILD_BUG_ON((offsetof(struct rte_ring, cons) &
			  RTE_CACHE_LINE_MASK) != 0);
	RTE_BUILD_BUG_ON((offsetof(struct rte_ring, prod) &
			  RTE_CACHE_LINE_MASK) != 0);

	/* init the ring structure */
	memset(r, 0, sizeof(*r));
	ret = strlcpy(r->name, name, sizeof(r->name));
	if (ret < 0 || ret >= (int)sizeof(r->name))
		return -ENAMETOOLONG;
	r->flags = flags;
	r->prod.single = (flags & RING_F_SP_ENQ) ? __IS_SP : __IS_MP;
	r->cons.single = (flags & RING_F_SC_DEQ) ? __IS_SC : __IS_MC;

	if (flags & RING_F_EXACT_SZ) {
		r->size = rte_align32pow2(count + 1);
		r->mask = r->size - 1;
		r->capacity = count;
	} else {
		if ((!POWEROF2(count)) || (count > RTE_RING_SZ_MASK)) {
			RTE_LOG(ERR, RING,
				"Requested size is invalid, must be power of 2, and not exceed the size limit %u\n",
				RTE_RING_SZ_MASK);
			return -EINVAL;
		}
		r->size = count;
		r->mask = count - 1;
		r->capacity = r->mask;
	}
	r->prod.head = r->cons.head = 0;
	r->prod.tail = r->cons.tail = 0;

	return 0;
}

/* create the ring */
struct rte_ring *
rte_ring_create(const char *name, unsigned count, int socket_id,
		unsigned flags)
{
	char mz_name[RTE_MEMZONE_NAMESIZE];
	struct rte_ring *r;
	struct rte_tailq_entry *te;
	const struct rte_memzone *mz;
	ssize_t ring_size;
	int mz_flags = 0;
	struct rte_ring_list* ring_list = NULL;
	const unsigned int requested_count = count;
	int ret;

	ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);

	/* for an exact size ring, round up from count to a power of two */
	if (flags & RING_F_EXACT_SZ)
		count = rte_align32pow2(count + 1);

	ring_size = rte_ring_get_memsize(count);
	if (ring_size < 0) {
		rte_errno = ring_size;
		return NULL;
	}

	ret = snprintf(mz_name, sizeof(mz_name), "%s%s",
		RTE_RING_MZ_PREFIX, name);
	if (ret < 0 || ret >= (int)sizeof(mz_name)) {
		rte_errno = ENAMETOOLONG;
		return NULL;
	}

	te = rte_zmalloc("RING_TAILQ_ENTRY", sizeof(*te), 0);
	if (te == NULL) {
		RTE_LOG(ERR, RING, "Cannot reserve memory for tailq\n");
		rte_errno = ENOMEM;
		return NULL;
	}

	rte_rwlock_write_lock(RTE_EAL_TAILQ_RWLOCK);

	/* reserve a memory zone for this ring. If we can't get rte_config or
	 * we are secondary process, the memzone_reserve function will set
	 * rte_errno for us appropriately - hence no check in this this function */
	mz = rte_memzone_reserve_aligned(mz_name, ring_size, socket_id,
					 mz_flags, __alignof__(*r));
	if (mz != NULL) {
		r = mz->addr;
		/* no need to check return value here, we already checked the
		 * arguments above */
		rte_ring_init(r, name, requested_count, flags);

		te->data = (void *) r;
		r->memzone = mz;

		TAILQ_INSERT_TAIL(ring_list, te, next);
	} else {
		r = NULL;
		RTE_LOG(ERR, RING, "Cannot reserve memory\n");
		rte_free(te);
	}
	rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);

	return r;
}

/* free the ring */
void
rte_ring_free(struct rte_ring *r)
{
	struct rte_ring_list *ring_list = NULL;
	struct rte_tailq_entry *te;

	if (r == NULL)
		return;

	/*
	 * Ring was not created with rte_ring_create,
	 * therefore, there is no memzone to free.
	 */
	if (r->memzone == NULL) {
		RTE_LOG(ERR, RING,
			"Cannot free ring, not created with rte_ring_create()\n");
		return;
	}

	if (rte_memzone_free(r->memzone) != 0) {
		RTE_LOG(ERR, RING, "Cannot free memory\n");
		return;
	}

	ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
	rte_rwlock_write_lock(RTE_EAL_TAILQ_RWLOCK);

	/* find out tailq entry */
	TAILQ_FOREACH(te, ring_list, next) {
		if (te->data == (void *) r)
			break;
	}

	if (te == NULL) {
		rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);
		return;
	}

	TAILQ_REMOVE(ring_list, te, next);

	rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);

	rte_free(te);
}

/* dump the status of the ring on the console */
void
rte_ring_dump(FILE *f, const struct rte_ring *r)
{
	fprintf(f, "ring <%s>@%p\n", r->name, r);
	fprintf(f, "  flags=%x\n", r->flags);
	fprintf(f, "  size=%"PRIu32"\n", r->size);
	fprintf(f, "  capacity=%"PRIu32"\n", r->capacity);
	fprintf(f, "  ct=%"PRIu32"\n", r->cons.tail);
	fprintf(f, "  ch=%"PRIu32"\n", r->cons.head);
	fprintf(f, "  pt=%"PRIu32"\n", r->prod.tail);
	fprintf(f, "  ph=%"PRIu32"\n", r->prod.head);
	fprintf(f, "  used=%u\n", rte_ring_count(r));
	fprintf(f, "  avail=%u\n", rte_ring_free_count(r));
}

/* dump the status of all rings on the console */
void
rte_ring_list_dump(FILE *f)
{
	const struct rte_tailq_entry *te;
	struct rte_ring_list *ring_list;

	ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);

	rte_rwlock_read_lock(RTE_EAL_TAILQ_RWLOCK);

	TAILQ_FOREACH(te, ring_list, next) {
		rte_ring_dump(f, (struct rte_ring *) te->data);
	}

	rte_rwlock_read_unlock(RTE_EAL_TAILQ_RWLOCK);
}

/* search a ring from its name */
struct rte_ring *
rte_ring_lookup(const char *name)
{
	struct rte_tailq_entry *te;
	struct rte_ring *r = NULL;
	struct rte_ring_list *ring_list;

	ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);

	rte_rwlock_read_lock(RTE_EAL_TAILQ_RWLOCK);

	TAILQ_FOREACH(te, ring_list, next) {
		r = (struct rte_ring *) te->data;
		if (strncmp(name, r->name, RTE_RING_NAMESIZE) == 0)
			break;
	}

	rte_rwlock_read_unlock(RTE_EAL_TAILQ_RWLOCK);

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

	return r;
}
Commit	Line	Data
9f95a23c	1	/* SPDX-License-Identifier: BSD-3-Clause
7c673cae	2	*
9f95a23c	3	* Copyright (c) 2010-2015 Intel Corporation
7c673cae FG	4	* Copyright (c) 2007,2008 Kip Macy kmacy@freebsd.org
7c673cae FG	5	* All rights reserved.
9f95a23c TL	6	* Derived from FreeBSD's bufring.h
	7	* Used as BSD-3 Licensed with permission from Kip Macy.
	8	*/
7c673cae FG	9
	10	#include <stdio.h>
	11	#include <stdarg.h>
	12	#include <string.h>
	13	#include <stdint.h>
	14	#include <inttypes.h>
	15	#include <errno.h>
	16	#include <sys/queue.h>
	17
	18	#include <rte_common.h>
	19	#include <rte_log.h>
	20	#include <rte_memory.h>
	21	#include <rte_memzone.h>
	22	#include <rte_malloc.h>
	23	#include <rte_launch.h>
	24	#include <rte_eal.h>
	25	#include <rte_eal_memconfig.h>
	26	#include <rte_atomic.h>
	27	#include <rte_per_lcore.h>
	28	#include <rte_lcore.h>
	29	#include <rte_branch_prediction.h>
	30	#include <rte_errno.h>
	31	#include <rte_string_fns.h>
	32	#include <rte_spinlock.h>
	33
	34	#include "rte_ring.h"
	35
	36	TAILQ_HEAD(rte_ring_list, rte_tailq_entry);
	37
	38	static struct rte_tailq_elem rte_ring_tailq = {
	39	.name = RTE_TAILQ_RING_NAME,
	40	};
	41	EAL_REGISTER_TAILQ(rte_ring_tailq)
	42
	43	/* true if x is a power of 2 */
	44	#define POWEROF2(x) ((((x)-1) & (x)) == 0)
	45
	46	/* return the size of memory occupied by a ring */
	47	ssize_t
	48	rte_ring_get_memsize(unsigned count)
	49	{
	50	ssize_t sz;
	51
	52	/* count must be a power of 2 */
	53	if ((!POWEROF2(count)) \|\| (count > RTE_RING_SZ_MASK )) {
	54	RTE_LOG(ERR, RING,
	55	"Requested size is invalid, must be power of 2, and "
	56	"do not exceed the size limit %u\n", RTE_RING_SZ_MASK);
	57	return -EINVAL;
	58	}
	59
	60	sz = sizeof(struct rte_ring) + count * sizeof(void *);
	61	sz = RTE_ALIGN(sz, RTE_CACHE_LINE_SIZE);
	62	return sz;
	63	}
	64
	65	int
	66	rte_ring_init(struct rte_ring r, const char name, unsigned count,
	67	unsigned flags)
	68	{
	69	int ret;
	70
	71	/* compilation-time checks */
	72	RTE_BUILD_BUG_ON((sizeof(struct rte_ring) &
73	RTE_CACHE_LINE_MASK) != 0);
7c673cae FG	74	RTE_BUILD_BUG_ON((offsetof(struct rte_ring, cons) &
7c673cae FG	75	RTE_CACHE_LINE_MASK) != 0);
7c673cae FG	76	RTE_BUILD_BUG_ON((offsetof(struct rte_ring, prod) &
7c673cae FG	77	RTE_CACHE_LINE_MASK) != 0);
7c673cae FG	78
	79	/* init the ring structure */
	80	memset(r, 0, sizeof(*r));
9f95a23c	81	ret = strlcpy(r->name, name, sizeof(r->name));
7c673cae FG	82	if (ret < 0 \|\| ret >= (int)sizeof(r->name))
	83	return -ENAMETOOLONG;
	84	r->flags = flags;
11fdf7f2 TL	85	r->prod.single = (flags & RING_F_SP_ENQ) ? __IS_SP : __IS_MP;
11fdf7f2 TL	86	r->cons.single = (flags & RING_F_SC_DEQ) ? __IS_SC : __IS_MC;
9f95a23c TL	87
	88	if (flags & RING_F_EXACT_SZ) {
	89	r->size = rte_align32pow2(count + 1);
	90	r->mask = r->size - 1;
	91	r->capacity = count;
	92	} else {
	93	if ((!POWEROF2(count)) \|\| (count > RTE_RING_SZ_MASK)) {
	94	RTE_LOG(ERR, RING,
	95	"Requested size is invalid, must be power of 2, and not exceed the size limit %u\n",
	96	RTE_RING_SZ_MASK);
	97	return -EINVAL;
	98	}
	99	r->size = count;
	100	r->mask = count - 1;
	101	r->capacity = r->mask;
	102	}
7c673cae FG	103	r->prod.head = r->cons.head = 0;
	104	r->prod.tail = r->cons.tail = 0;
	105
	106	return 0;
	107	}
	108
	109	/* create the ring */
	110	struct rte_ring *
	111	rte_ring_create(const char *name, unsigned count, int socket_id,
	112	unsigned flags)
	113	{
	114	char mz_name[RTE_MEMZONE_NAMESIZE];
	115	struct rte_ring *r;
	116	struct rte_tailq_entry *te;
	117	const struct rte_memzone *mz;
	118	ssize_t ring_size;
	119	int mz_flags = 0;
	120	struct rte_ring_list* ring_list = NULL;
9f95a23c	121	const unsigned int requested_count = count;
7c673cae FG	122	int ret;
	123
	124	ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
	125
9f95a23c TL	126	/* for an exact size ring, round up from count to a power of two */
	127	if (flags & RING_F_EXACT_SZ)
	128	count = rte_align32pow2(count + 1);
	129
7c673cae FG	130	ring_size = rte_ring_get_memsize(count);
	131	if (ring_size < 0) {
	132	rte_errno = ring_size;
	133	return NULL;
	134	}
	135
	136	ret = snprintf(mz_name, sizeof(mz_name), "%s%s",
	137	RTE_RING_MZ_PREFIX, name);
	138	if (ret < 0 \|\| ret >= (int)sizeof(mz_name)) {
	139	rte_errno = ENAMETOOLONG;
	140	return NULL;
	141	}
	142
	143	te = rte_zmalloc("RING_TAILQ_ENTRY", sizeof(*te), 0);
	144	if (te == NULL) {
	145	RTE_LOG(ERR, RING, "Cannot reserve memory for tailq\n");
	146	rte_errno = ENOMEM;
	147	return NULL;
	148	}
	149
	150	rte_rwlock_write_lock(RTE_EAL_TAILQ_RWLOCK);
	151
	152	/* reserve a memory zone for this ring. If we can't get rte_config or
	153	* we are secondary process, the memzone_reserve function will set
	154	* rte_errno for us appropriately - hence no check in this this function */
9f95a23c TL	155	mz = rte_memzone_reserve_aligned(mz_name, ring_size, socket_id,
9f95a23c TL	156	mz_flags, __alignof__(*r));
7c673cae FG	157	if (mz != NULL) {
	158	r = mz->addr;
	159	/* no need to check return value here, we already checked the
	160	* arguments above */
9f95a23c	161	rte_ring_init(r, name, requested_count, flags);
7c673cae FG	162
	163	te->data = (void *) r;
	164	r->memzone = mz;
	165
	166	TAILQ_INSERT_TAIL(ring_list, te, next);
	167	} else {
	168	r = NULL;
	169	RTE_LOG(ERR, RING, "Cannot reserve memory\n");
	170	rte_free(te);
	171	}
	172	rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);
	173
	174	return r;
	175	}
	176
	177	/* free the ring */
	178	void
	179	rte_ring_free(struct rte_ring *r)
	180	{
	181	struct rte_ring_list *ring_list = NULL;
	182	struct rte_tailq_entry *te;
	183
	184	if (r == NULL)
	185	return;
	186
	187	/*
	188	* Ring was not created with rte_ring_create,
	189	* therefore, there is no memzone to free.
	190	*/
	191	if (r->memzone == NULL) {
9f95a23c TL	192	RTE_LOG(ERR, RING,
9f95a23c TL	193	"Cannot free ring, not created with rte_ring_create()\n");
7c673cae FG	194	return;
	195	}
	196
	197	if (rte_memzone_free(r->memzone) != 0) {
	198	RTE_LOG(ERR, RING, "Cannot free memory\n");
	199	return;
	200	}
	201
	202	ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
	203	rte_rwlock_write_lock(RTE_EAL_TAILQ_RWLOCK);
	204
	205	/* find out tailq entry */
	206	TAILQ_FOREACH(te, ring_list, next) {
	207	if (te->data == (void *) r)
	208	break;
	209	}
	210
	211	if (te == NULL) {
	212	rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);
	213	return;
	214	}
	215
	216	TAILQ_REMOVE(ring_list, te, next);
	217
	218	rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);
	219
	220	rte_free(te);
	221	}
	222
7c673cae FG	223	/* dump the status of the ring on the console */
	224	void
	225	rte_ring_dump(FILE f, const struct rte_ring r)
	226	{
7c673cae FG	227	fprintf(f, "ring <%s>@%p\n", r->name, r);
7c673cae FG	228	fprintf(f, " flags=%x\n", r->flags);
11fdf7f2	229	fprintf(f, " size=%"PRIu32"\n", r->size);
9f95a23c	230	fprintf(f, " capacity=%"PRIu32"\n", r->capacity);
7c673cae FG	231	fprintf(f, " ct=%"PRIu32"\n", r->cons.tail);
	232	fprintf(f, " ch=%"PRIu32"\n", r->cons.head);
	233	fprintf(f, " pt=%"PRIu32"\n", r->prod.tail);
	234	fprintf(f, " ph=%"PRIu32"\n", r->prod.head);
	235	fprintf(f, " used=%u\n", rte_ring_count(r));
	236	fprintf(f, " avail=%u\n", rte_ring_free_count(r));
7c673cae FG	237	}
	238
	239	/* dump the status of all rings on the console */
	240	void
	241	rte_ring_list_dump(FILE *f)
	242	{
	243	const struct rte_tailq_entry *te;
	244	struct rte_ring_list *ring_list;
	245
	246	ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
	247
	248	rte_rwlock_read_lock(RTE_EAL_TAILQ_RWLOCK);
	249
	250	TAILQ_FOREACH(te, ring_list, next) {
	251	rte_ring_dump(f, (struct rte_ring *) te->data);
	252	}
	253
	254	rte_rwlock_read_unlock(RTE_EAL_TAILQ_RWLOCK);
	255	}
	256
	257	/* search a ring from its name */
	258	struct rte_ring *
	259	rte_ring_lookup(const char *name)
	260	{
	261	struct rte_tailq_entry *te;
	262	struct rte_ring *r = NULL;
	263	struct rte_ring_list *ring_list;
	264
	265	ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
	266
	267	rte_rwlock_read_lock(RTE_EAL_TAILQ_RWLOCK);
	268
	269	TAILQ_FOREACH(te, ring_list, next) {
	270	r = (struct rte_ring *) te->data;
	271	if (strncmp(name, r->name, RTE_RING_NAMESIZE) == 0)
	272	break;
	273	}
	274
	275	rte_rwlock_read_unlock(RTE_EAL_TAILQ_RWLOCK);
	276
	277	if (te == NULL) {
	278	rte_errno = ENOENT;
	279	return NULL;
	280	}
	281
	282	return r;
	283	}