From patchwork Wed Jun 7 13:36:18 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Bruce Richardson X-Patchwork-Id: 25107 Return-Path: X-Original-To: patchwork@dpdk.org Delivered-To: patchwork@dpdk.org Received: from [92.243.14.124] (localhost [IPv6:::1]) by dpdk.org (Postfix) with ESMTP id DFCD43251; Wed, 7 Jun 2017 16:05:40 +0200 (CEST) Received: from mga07.intel.com (mga07.intel.com [134.134.136.100]) by dpdk.org (Postfix) with ESMTP id 9606D2C2F for ; Wed, 7 Jun 2017 16:05:38 +0200 (CEST) Received: from fmsmga001.fm.intel.com ([10.253.24.23]) by orsmga105.jf.intel.com with ESMTP; 07 Jun 2017 07:05:38 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos; i="5.39,311,1493708400"; d="scan'208"; a="1157667633" Received: from silpixa00399126.ir.intel.com (HELO silpixa00399126.ger.corp.intel.com) ([10.237.223.223]) by fmsmga001.fm.intel.com with ESMTP; 07 Jun 2017 07:05:36 -0700 From: Bruce Richardson To: dev@dpdk.org Cc: olivier.matz@6wind.com, jerin.jacob@caviumnetworks.com, Bruce Richardson Date: Wed, 7 Jun 2017 14:36:18 +0100 Message-Id: <20170607133620.275801-4-bruce.richardson@intel.com> X-Mailer: git-send-email 2.9.4 In-Reply-To: <20170607133620.275801-1-bruce.richardson@intel.com> References: <20170607133620.275801-1-bruce.richardson@intel.com> Subject: [dpdk-dev] [PATCH 3/5] eventdev: add ring structure for events X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" Add in a new rte_event_ring structure type and functions to allow events to be passed core to core. This is needed because the standard rte_ring type only works on pointers, while for events, we want to copy the entire, 16B events themselves - not just pointers to them. The code makes extensive use of the functions already defined in rte_ring.h Signed-off-by: Bruce Richardson --- lib/Makefile | 2 +- lib/librte_eventdev/Makefile | 2 + lib/librte_eventdev/rte_event_ring.c | 207 ++++++++++++++++++ lib/librte_eventdev/rte_event_ring.h | 312 +++++++++++++++++++++++++++ lib/librte_eventdev/rte_eventdev_version.map | 9 + 5 files changed, 531 insertions(+), 1 deletion(-) create mode 100644 lib/librte_eventdev/rte_event_ring.c create mode 100644 lib/librte_eventdev/rte_event_ring.h diff --git a/lib/Makefile b/lib/Makefile index 07e1fd0..aef584e 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -52,7 +52,7 @@ DIRS-$(CONFIG_RTE_LIBRTE_CRYPTODEV) += librte_cryptodev DEPDIRS-librte_cryptodev := librte_eal librte_mempool librte_ring librte_mbuf DEPDIRS-librte_cryptodev += librte_kvargs DIRS-$(CONFIG_RTE_LIBRTE_EVENTDEV) += librte_eventdev -DEPDIRS-librte_eventdev := librte_eal +DEPDIRS-librte_eventdev := librte_eal librte_ring DIRS-$(CONFIG_RTE_LIBRTE_VHOST) += librte_vhost DEPDIRS-librte_vhost := librte_eal librte_mempool librte_mbuf librte_ether DIRS-$(CONFIG_RTE_LIBRTE_HASH) += librte_hash diff --git a/lib/librte_eventdev/Makefile b/lib/librte_eventdev/Makefile index e06346a..0313277 100644 --- a/lib/librte_eventdev/Makefile +++ b/lib/librte_eventdev/Makefile @@ -42,10 +42,12 @@ CFLAGS += $(WERROR_FLAGS) # library source files SRCS-y += rte_eventdev.c +SRCS-y += rte_event_ring.c # export include files SYMLINK-y-include += rte_eventdev.h SYMLINK-y-include += rte_eventdev_pmd.h +SYMLINK-y-include += rte_event_ring.h # versioning export map EXPORT_MAP := rte_eventdev_version.map diff --git a/lib/librte_eventdev/rte_event_ring.c b/lib/librte_eventdev/rte_event_ring.c new file mode 100644 index 0000000..b14c212 --- /dev/null +++ b/lib/librte_eventdev/rte_event_ring.c @@ -0,0 +1,207 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2017 Intel Corporation. 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 +#include + +#include +#include +#include +#include +#include "rte_event_ring.h" + +TAILQ_HEAD(rte_event_ring_list, rte_tailq_entry); + +static struct rte_tailq_elem rte_event_ring_tailq = { + .name = RTE_TAILQ_EVENT_RING_NAME, +}; +EAL_REGISTER_TAILQ(rte_event_ring_tailq) + +int +rte_event_ring_init(struct rte_event_ring *r, const char *name, + unsigned int count, unsigned int flags) +{ + /* compilation-time checks */ + RTE_BUILD_BUG_ON((sizeof(struct rte_event_ring) & + RTE_CACHE_LINE_MASK) != 0); + + /* init the ring structure */ + return rte_ring_init(&r->r, name, count, flags); +} + +/* create the ring */ +struct rte_event_ring * +rte_event_ring_create(const char *name, unsigned int count, int socket_id, + unsigned int flags) +{ + char mz_name[RTE_MEMZONE_NAMESIZE]; + struct rte_event_ring *r; + struct rte_tailq_entry *te; + const struct rte_memzone *mz; + ssize_t ring_size; + int mz_flags = 0; + struct rte_event_ring_list *ring_list = NULL; + const unsigned int requested_count = count; + int ret; + + ring_list = RTE_TAILQ_CAST(rte_event_ring_tailq.head, + rte_event_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); + else if (!rte_is_power_of_2(count)) { + rte_errno = EINVAL; + return NULL; + } + + ring_size = sizeof(*r) + (count * sizeof(struct rte_event)); + + 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(mz_name, ring_size, socket_id, mz_flags); + if (mz != NULL) { + r = mz->addr; + /* + * no need to check return value here, we already checked the + * arguments above + */ + rte_event_ring_init(r, name, requested_count, flags); + + te->data = (void *) r; + 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; +} + + +struct rte_event_ring * +rte_event_ring_lookup(const char *name) +{ + struct rte_tailq_entry *te; + struct rte_event_ring *r = NULL; + struct rte_event_ring_list *ring_list; + + ring_list = RTE_TAILQ_CAST(rte_event_ring_tailq.head, + rte_event_ring_list); + + rte_rwlock_read_lock(RTE_EAL_TAILQ_RWLOCK); + + TAILQ_FOREACH(te, ring_list, next) { + r = (struct rte_event_ring *) te->data; + if (strncmp(name, r->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; +} + +/* free the ring */ +void +rte_event_ring_free(struct rte_event_ring *r) +{ + struct rte_event_ring_list *ring_list = NULL; + struct rte_tailq_entry *te; + + if (r == NULL) + return; + + /* + * Ring was not created with rte_event_ring_create, + * therefore, there is no memzone to free. + */ + if (r->r.memzone == NULL) { + RTE_LOG(ERR, RING, + "Cannot free ring (not created with rte_event_ring_create()"); + return; + } + + if (rte_memzone_free(r->r.memzone) != 0) { + RTE_LOG(ERR, RING, "Cannot free memory\n"); + return; + } + + ring_list = RTE_TAILQ_CAST(rte_event_ring_tailq.head, + rte_event_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); +} diff --git a/lib/librte_eventdev/rte_event_ring.h b/lib/librte_eventdev/rte_event_ring.h new file mode 100644 index 0000000..f7eae58 --- /dev/null +++ b/lib/librte_eventdev/rte_event_ring.h @@ -0,0 +1,312 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2016-2017 Intel Corporation. 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. + */ + +/** + * @file + * RTE Event Ring + * + * This provides a ring implementation for passing rte_event structures + * from one core to another. + */ + +#ifndef _RTE_EVENT_RING_ +#define _RTE_EVENT_RING_ + +#include + +#include +#include +#include +#include +#include "rte_eventdev.h" + +#define RTE_TAILQ_EVENT_RING_NAME "RTE_EVENT_RING" + +#ifndef force_inline +#define force_inline inline __attribute__((always_inline)) +#endif + +/** + * Generic ring structure for passing rte_event objects from core to core. + * + * Based on the primitives given in the rte_ring library. Designed to be + * used inside software eventdev implementations and by applications + * directly as needed. + */ +struct rte_event_ring { + struct rte_ring r; +}; + +/** + * Returns the number of events in the ring + * + * @param r + * pointer to the event ring + * @return + * the number of events in the ring + */ +static force_inline unsigned int +rte_event_ring_count(const struct rte_event_ring *r) +{ + return rte_ring_count(&r->r); +} + +/** + * Returns the amount of free space in the ring + * + * @param r + * pointer to the event ring + * @return + * the number of free slots in the ring, i.e. the number of events that + * can be successfully enqueued before dequeue must be called + */ +static force_inline unsigned int +rte_event_ring_free_count(const struct rte_event_ring *r) +{ + return rte_ring_free_count(&r->r); +} + +/** + * Enqueue a set of events onto a ring + * + * Note: this API enqueues by copying the events themselves onto the ring, + * rather than just placing a pointer to each event onto the ring. This + * means that statically-allocated events can safely be enqueued by this + * API. + * + * @param r + * pointer to the event ring + * @param events + * pointer to an array of struct rte_event objects + * @param n + * number of events in the array to enqueue + * @param free_space + * if non-null, is updated to indicate the amount of free space in the + * ring once the enqueue has completed. + * @return + * the number of elements, n', enqueued to the ring, 0 <= n' <= n + */ +static force_inline unsigned int +rte_event_ring_enqueue_burst(struct rte_event_ring *r, + const struct rte_event *events, + unsigned int n, uint16_t *free_space) +{ + uint32_t prod_head, prod_next; + uint32_t free_entries; + + n = __rte_ring_move_prod_head(&r->r, r->r.prod.single, n, + RTE_RING_QUEUE_VARIABLE, + &prod_head, &prod_next, &free_entries); + if (n == 0) + goto end; + + ENQUEUE_PTRS(&r->r, &r[1], prod_head, events, n, struct rte_event); + rte_smp_wmb(); + + update_tail(&r->r.prod, prod_head, prod_next, 1); +end: + if (free_space != NULL) + *free_space = free_entries - n; + return n; +} + +/** + * Dequeue a set of events from a ring + * + * Note: this API does not work with pointers to events, rather it copies + * the events themselves to the destination ``events`` buffer. + * + * @param r + * pointer to the event ring + * @param events + * pointer to an array to hold the struct rte_event objects + * @param n + * number of events that can be held in the ``events`` array + * @param available + * if non-null, is updated to indicate the number of events remaining in + * the ring once the dequeue has completed + * @return + * the number of elements, n', dequeued from the ring, 0 <= n' <= n + */ +static force_inline unsigned int +rte_event_ring_dequeue_burst(struct rte_event_ring *r, + struct rte_event *events, + unsigned int n, uint16_t *available) +{ + uint32_t cons_head, cons_next; + uint32_t entries; + + n = __rte_ring_move_cons_head(&r->r, r->r.cons.single, n, + RTE_RING_QUEUE_VARIABLE, + &cons_head, &cons_next, &entries); + if (n == 0) + goto end; + + DEQUEUE_PTRS(&r->r, &r[1], cons_head, events, n, struct rte_event); + rte_smp_rmb(); + + update_tail(&r->r.cons, cons_head, cons_next, 1); + +end: + if (available != NULL) + *available = entries - n; + return n; +} + +/* + * Initializes an already-allocated ring structure + * + * @param r + * pointer to the ring memory to be initialized + * @param name + * name to be given to the ring + * @param count + * the number of elements to be stored in the ring. If the flag + * ``RING_F_EXACT_SZ`` is not set, this must be a power of 2, and the actual + * usable space in the ring will be ``count - 1`` entries. If the flag + * ``RING_F_EXACT_SZ`` is set, the this can be any value up to the ring size + * limit - 1, and the usable space will be exactly that requested. + * @param flags + * An OR of the following: + * - RING_F_SP_ENQ: If this flag is set, the default behavior when + * using ``rte_ring_enqueue()`` or ``rte_ring_enqueue_bulk()`` + * is "single-producer". Otherwise, it is "multi-producers". + * - RING_F_SC_DEQ: If this flag is set, the default behavior when + * using ``rte_ring_dequeue()`` or ``rte_ring_dequeue_bulk()`` + * is "single-consumer". Otherwise, it is "multi-consumers". + * - RING_F_EXACT_SZ: If this flag is set, the ``count`` parameter is to + * be taken as the exact usable size of the ring, and as such does not + * need to be a power of 2. The underlying ring memory should be a + * power-of-2 size greater than the count value. + * @return + * 0 on success, or a negative value on error. + */ +int +rte_event_ring_init(struct rte_event_ring *r, const char *name, + unsigned int count, unsigned int flags); + +/* + * Create an event ring structure + * + * This function allocates memory and initializes an event ring inside that + * memory. + * + * @param name + * name to be given to the ring + * @param count + * the number of elements to be stored in the ring. If the flag + * ``RING_F_EXACT_SZ`` is not set, this must be a power of 2, and the actual + * usable space in the ring will be ``count - 1`` entries. If the flag + * ``RING_F_EXACT_SZ`` is set, the this can be any value up to the ring size + * limit - 1, and the usable space will be exactly that requested. + * @param socket_id + * The *socket_id* argument is the socket identifier in case of + * NUMA. The value can be *SOCKET_ID_ANY* if there is no NUMA + * constraint for the reserved zone. + * @param flags + * An OR of the following: + * - RING_F_SP_ENQ: If this flag is set, the default behavior when + * using ``rte_ring_enqueue()`` or ``rte_ring_enqueue_bulk()`` + * is "single-producer". Otherwise, it is "multi-producers". + * - RING_F_SC_DEQ: If this flag is set, the default behavior when + * using ``rte_ring_dequeue()`` or ``rte_ring_dequeue_bulk()`` + * is "single-consumer". Otherwise, it is "multi-consumers". + * - RING_F_EXACT_SZ: If this flag is set, the ``count`` parameter is to + * be taken as the exact usable size of the ring, and as such does not + * need to be a power of 2. The underlying ring memory should be a + * power-of-2 size greater than the count value. + * @return + * On success, the pointer to the new allocated ring. NULL on error with + * rte_errno set appropriately. Possible errno values include: + * - E_RTE_NO_CONFIG - function could not get pointer to rte_config structure + * - E_RTE_SECONDARY - function was called from a secondary process instance + * - EINVAL - count provided is not a power of 2 + * - ENOSPC - the maximum number of memzones has already been allocated + * - EEXIST - a memzone with the same name already exists + * - ENOMEM - no appropriate memory area found in which to create memzone + */ +struct rte_event_ring * +rte_event_ring_create(const char *name, unsigned int count, int socket_id, + unsigned int flags); + +/** + * Search for an event ring based on its name + * + * @param name + * The name of the ring. + * @return + * The pointer to the ring matching the name, or NULL if not found, + * with rte_errno set appropriately. Possible rte_errno values include: + * - ENOENT - required entry not available to return. + */ +struct rte_event_ring * +rte_event_ring_lookup(const char *name); + +/** + * De-allocate all memory used by the ring. + * + * @param r + * Ring to free + */ +void +rte_event_ring_free(struct rte_event_ring *r); + +/** + * Return the size of the event ring. + * + * @param r + * A pointer to the ring structure. + * @return + * The size of the data store used by the ring. + * NOTE: this is not the same as the usable space in the ring. To query that + * use ``rte_ring_get_capacity()``. + */ +static inline unsigned int +rte_event_ring_get_size(const struct rte_event_ring *r) +{ + return rte_ring_get_size(&r->r); +} + +/** + * Return the number of elements which can be stored in the event ring. + * + * @param r + * A pointer to the ring structure. + * @return + * The usable size of the ring. + */ +static inline unsigned int +rte_event_ring_get_capacity(const struct rte_event_ring *r) +{ + return rte_ring_get_capacity(&r->r); +} +#endif diff --git a/lib/librte_eventdev/rte_eventdev_version.map b/lib/librte_eventdev/rte_eventdev_version.map index 1fa6b33..4c48e5f 100644 --- a/lib/librte_eventdev/rte_eventdev_version.map +++ b/lib/librte_eventdev/rte_eventdev_version.map @@ -42,3 +42,12 @@ DPDK_17.05 { local: *; }; + +DPDK_17.08 { + global: + + rte_event_ring_create; + rte_event_ring_free; + rte_event_ring_init; + rte_event_ring_lookup; +} DPDK_17.05;