From 97f17497d162afdb82c8704bf097f0fee3724b2e Mon Sep 17 00:00:00 2001 From: "C.J. Collier" Date: Tue, 14 Jun 2016 07:50:17 -0700 Subject: Imported Upstream version 16.04 Change-Id: I77eadcd8538a9122e4773cbe55b24033dc451757 Signed-off-by: C.J. Collier --- drivers/net/mlx5/mlx5_txq.c | 605 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 605 insertions(+) create mode 100644 drivers/net/mlx5/mlx5_txq.c (limited to 'drivers/net/mlx5/mlx5_txq.c') diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c new file mode 100644 index 00000000..31ce53ad --- /dev/null +++ b/drivers/net/mlx5/mlx5_txq.c @@ -0,0 +1,605 @@ +/*- + * BSD LICENSE + * + * Copyright 2015 6WIND S.A. + * Copyright 2015 Mellanox. + * + * 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 6WIND S.A. 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 + +/* Verbs header. */ +/* ISO C doesn't support unnamed structs/unions, disabling -pedantic. */ +#ifdef PEDANTIC +#pragma GCC diagnostic ignored "-pedantic" +#endif +#include +#ifdef PEDANTIC +#pragma GCC diagnostic error "-pedantic" +#endif + +/* DPDK headers don't like -pedantic. */ +#ifdef PEDANTIC +#pragma GCC diagnostic ignored "-pedantic" +#endif +#include +#include +#include +#include +#ifdef PEDANTIC +#pragma GCC diagnostic error "-pedantic" +#endif + +#include "mlx5_utils.h" +#include "mlx5.h" +#include "mlx5_rxtx.h" +#include "mlx5_autoconf.h" +#include "mlx5_defs.h" + +/** + * Allocate TX queue elements. + * + * @param txq + * Pointer to TX queue structure. + * @param elts_n + * Number of elements to allocate. + * + * @return + * 0 on success, errno value on failure. + */ +static int +txq_alloc_elts(struct txq *txq, unsigned int elts_n) +{ + unsigned int i; + struct txq_elt (*elts)[elts_n] = + rte_calloc_socket("TXQ", 1, sizeof(*elts), 0, txq->socket); + linear_t (*elts_linear)[elts_n] = + rte_calloc_socket("TXQ", 1, sizeof(*elts_linear), 0, + txq->socket); + struct ibv_mr *mr_linear = NULL; + int ret = 0; + + if ((elts == NULL) || (elts_linear == NULL)) { + ERROR("%p: can't allocate packets array", (void *)txq); + ret = ENOMEM; + goto error; + } + mr_linear = + ibv_reg_mr(txq->priv->pd, elts_linear, sizeof(*elts_linear), + (IBV_ACCESS_LOCAL_WRITE | IBV_ACCESS_REMOTE_WRITE)); + if (mr_linear == NULL) { + ERROR("%p: unable to configure MR, ibv_reg_mr() failed", + (void *)txq); + ret = EINVAL; + goto error; + } + for (i = 0; (i != elts_n); ++i) { + struct txq_elt *elt = &(*elts)[i]; + + elt->buf = NULL; + } + DEBUG("%p: allocated and configured %u WRs", (void *)txq, elts_n); + txq->elts_n = elts_n; + txq->elts = elts; + txq->elts_head = 0; + txq->elts_tail = 0; + txq->elts_comp = 0; + /* Request send completion every MLX5_PMD_TX_PER_COMP_REQ packets or + * at least 4 times per ring. */ + txq->elts_comp_cd_init = + ((MLX5_PMD_TX_PER_COMP_REQ < (elts_n / 4)) ? + MLX5_PMD_TX_PER_COMP_REQ : (elts_n / 4)); + txq->elts_comp_cd = txq->elts_comp_cd_init; + txq->elts_linear = elts_linear; + txq->mr_linear = mr_linear; + assert(ret == 0); + return 0; +error: + if (mr_linear != NULL) + claim_zero(ibv_dereg_mr(mr_linear)); + + rte_free(elts_linear); + rte_free(elts); + + DEBUG("%p: failed, freed everything", (void *)txq); + assert(ret > 0); + return ret; +} + +/** + * Free TX queue elements. + * + * @param txq + * Pointer to TX queue structure. + */ +static void +txq_free_elts(struct txq *txq) +{ + unsigned int elts_n = txq->elts_n; + unsigned int elts_head = txq->elts_head; + unsigned int elts_tail = txq->elts_tail; + struct txq_elt (*elts)[elts_n] = txq->elts; + linear_t (*elts_linear)[elts_n] = txq->elts_linear; + struct ibv_mr *mr_linear = txq->mr_linear; + + DEBUG("%p: freeing WRs", (void *)txq); + txq->elts_n = 0; + txq->elts_head = 0; + txq->elts_tail = 0; + txq->elts_comp = 0; + txq->elts_comp_cd = 0; + txq->elts_comp_cd_init = 0; + txq->elts = NULL; + txq->elts_linear = NULL; + txq->mr_linear = NULL; + if (mr_linear != NULL) + claim_zero(ibv_dereg_mr(mr_linear)); + + rte_free(elts_linear); + if (elts == NULL) + return; + while (elts_tail != elts_head) { + struct txq_elt *elt = &(*elts)[elts_tail]; + + assert(elt->buf != NULL); + rte_pktmbuf_free(elt->buf); +#ifndef NDEBUG + /* Poisoning. */ + memset(elt, 0x77, sizeof(*elt)); +#endif + if (++elts_tail == elts_n) + elts_tail = 0; + } + rte_free(elts); +} + +/** + * Clean up a TX queue. + * + * Destroy objects, free allocated memory and reset the structure for reuse. + * + * @param txq + * Pointer to TX queue structure. + */ +void +txq_cleanup(struct txq *txq) +{ + struct ibv_exp_release_intf_params params; + size_t i; + + DEBUG("cleaning up %p", (void *)txq); + txq_free_elts(txq); + txq->poll_cnt = NULL; +#if MLX5_PMD_MAX_INLINE > 0 + txq->send_pending_inline = NULL; +#endif + txq->send_flush = NULL; + if (txq->if_qp != NULL) { + assert(txq->priv != NULL); + assert(txq->priv->ctx != NULL); + assert(txq->qp != NULL); + params = (struct ibv_exp_release_intf_params){ + .comp_mask = 0, + }; + claim_zero(ibv_exp_release_intf(txq->priv->ctx, + txq->if_qp, + ¶ms)); + } + if (txq->if_cq != NULL) { + assert(txq->priv != NULL); + assert(txq->priv->ctx != NULL); + assert(txq->cq != NULL); + params = (struct ibv_exp_release_intf_params){ + .comp_mask = 0, + }; + claim_zero(ibv_exp_release_intf(txq->priv->ctx, + txq->if_cq, + ¶ms)); + } + if (txq->qp != NULL) + claim_zero(ibv_destroy_qp(txq->qp)); + if (txq->cq != NULL) + claim_zero(ibv_destroy_cq(txq->cq)); + if (txq->rd != NULL) { + struct ibv_exp_destroy_res_domain_attr attr = { + .comp_mask = 0, + }; + + assert(txq->priv != NULL); + assert(txq->priv->ctx != NULL); + claim_zero(ibv_exp_destroy_res_domain(txq->priv->ctx, + txq->rd, + &attr)); + } + for (i = 0; (i != RTE_DIM(txq->mp2mr)); ++i) { + if (txq->mp2mr[i].mp == NULL) + break; + assert(txq->mp2mr[i].mr != NULL); + claim_zero(ibv_dereg_mr(txq->mp2mr[i].mr)); + } + memset(txq, 0, sizeof(*txq)); +} + +/** + * Configure a TX queue. + * + * @param dev + * Pointer to Ethernet device structure. + * @param txq + * Pointer to TX queue structure. + * @param desc + * Number of descriptors to configure in queue. + * @param socket + * NUMA socket on which memory must be allocated. + * @param[in] conf + * Thresholds parameters. + * + * @return + * 0 on success, errno value on failure. + */ +int +txq_setup(struct rte_eth_dev *dev, struct txq *txq, uint16_t desc, + unsigned int socket, const struct rte_eth_txconf *conf) +{ + struct priv *priv = mlx5_get_priv(dev); + struct txq tmpl = { + .priv = priv, + .socket = socket + }; + union { + struct ibv_exp_query_intf_params params; + struct ibv_exp_qp_init_attr init; + struct ibv_exp_res_domain_init_attr rd; + struct ibv_exp_cq_init_attr cq; + struct ibv_exp_qp_attr mod; + } attr; + enum ibv_exp_query_intf_status status; + int ret = 0; + + (void)conf; /* Thresholds configuration (ignored). */ + if ((desc == 0) || (desc % MLX5_PMD_SGE_WR_N)) { + ERROR("%p: invalid number of TX descriptors (must be a" + " multiple of %d)", (void *)dev, MLX5_PMD_SGE_WR_N); + return EINVAL; + } + desc /= MLX5_PMD_SGE_WR_N; + /* MRs will be registered in mp2mr[] later. */ + attr.rd = (struct ibv_exp_res_domain_init_attr){ + .comp_mask = (IBV_EXP_RES_DOMAIN_THREAD_MODEL | + IBV_EXP_RES_DOMAIN_MSG_MODEL), + .thread_model = IBV_EXP_THREAD_SINGLE, + .msg_model = IBV_EXP_MSG_HIGH_BW, + }; + tmpl.rd = ibv_exp_create_res_domain(priv->ctx, &attr.rd); + if (tmpl.rd == NULL) { + ret = ENOMEM; + ERROR("%p: RD creation failure: %s", + (void *)dev, strerror(ret)); + goto error; + } + attr.cq = (struct ibv_exp_cq_init_attr){ + .comp_mask = IBV_EXP_CQ_INIT_ATTR_RES_DOMAIN, + .res_domain = tmpl.rd, + }; + tmpl.cq = ibv_exp_create_cq(priv->ctx, desc, NULL, NULL, 0, &attr.cq); + if (tmpl.cq == NULL) { + ret = ENOMEM; + ERROR("%p: CQ creation failure: %s", + (void *)dev, strerror(ret)); + goto error; + } + DEBUG("priv->device_attr.max_qp_wr is %d", + priv->device_attr.max_qp_wr); + DEBUG("priv->device_attr.max_sge is %d", + priv->device_attr.max_sge); + attr.init = (struct ibv_exp_qp_init_attr){ + /* CQ to be associated with the send queue. */ + .send_cq = tmpl.cq, + /* CQ to be associated with the receive queue. */ + .recv_cq = tmpl.cq, + .cap = { + /* Max number of outstanding WRs. */ + .max_send_wr = ((priv->device_attr.max_qp_wr < desc) ? + priv->device_attr.max_qp_wr : + desc), + /* Max number of scatter/gather elements in a WR. */ + .max_send_sge = ((priv->device_attr.max_sge < + MLX5_PMD_SGE_WR_N) ? + priv->device_attr.max_sge : + MLX5_PMD_SGE_WR_N), +#if MLX5_PMD_MAX_INLINE > 0 + .max_inline_data = MLX5_PMD_MAX_INLINE, +#endif + }, + .qp_type = IBV_QPT_RAW_PACKET, + /* Do *NOT* enable this, completions events are managed per + * TX burst. */ + .sq_sig_all = 0, + .pd = priv->pd, + .res_domain = tmpl.rd, + .comp_mask = (IBV_EXP_QP_INIT_ATTR_PD | + IBV_EXP_QP_INIT_ATTR_RES_DOMAIN), + }; + tmpl.qp = ibv_exp_create_qp(priv->ctx, &attr.init); + if (tmpl.qp == NULL) { + ret = (errno ? errno : EINVAL); + ERROR("%p: QP creation failure: %s", + (void *)dev, strerror(ret)); + goto error; + } +#if MLX5_PMD_MAX_INLINE > 0 + /* ibv_create_qp() updates this value. */ + tmpl.max_inline = attr.init.cap.max_inline_data; +#endif + attr.mod = (struct ibv_exp_qp_attr){ + /* Move the QP to this state. */ + .qp_state = IBV_QPS_INIT, + /* Primary port number. */ + .port_num = priv->port + }; + ret = ibv_exp_modify_qp(tmpl.qp, &attr.mod, + (IBV_EXP_QP_STATE | IBV_EXP_QP_PORT)); + if (ret) { + ERROR("%p: QP state to IBV_QPS_INIT failed: %s", + (void *)dev, strerror(ret)); + goto error; + } + ret = txq_alloc_elts(&tmpl, desc); + if (ret) { + ERROR("%p: TXQ allocation failed: %s", + (void *)dev, strerror(ret)); + goto error; + } + attr.mod = (struct ibv_exp_qp_attr){ + .qp_state = IBV_QPS_RTR + }; + ret = ibv_exp_modify_qp(tmpl.qp, &attr.mod, IBV_EXP_QP_STATE); + if (ret) { + ERROR("%p: QP state to IBV_QPS_RTR failed: %s", + (void *)dev, strerror(ret)); + goto error; + } + attr.mod.qp_state = IBV_QPS_RTS; + ret = ibv_exp_modify_qp(tmpl.qp, &attr.mod, IBV_EXP_QP_STATE); + if (ret) { + ERROR("%p: QP state to IBV_QPS_RTS failed: %s", + (void *)dev, strerror(ret)); + goto error; + } + attr.params = (struct ibv_exp_query_intf_params){ + .intf_scope = IBV_EXP_INTF_GLOBAL, + .intf = IBV_EXP_INTF_CQ, + .obj = tmpl.cq, + }; + tmpl.if_cq = ibv_exp_query_intf(priv->ctx, &attr.params, &status); + if (tmpl.if_cq == NULL) { + ret = EINVAL; + ERROR("%p: CQ interface family query failed with status %d", + (void *)dev, status); + goto error; + } + attr.params = (struct ibv_exp_query_intf_params){ + .intf_scope = IBV_EXP_INTF_GLOBAL, + .intf = IBV_EXP_INTF_QP_BURST, + .obj = tmpl.qp, +#ifdef HAVE_VERBS_VLAN_INSERTION + .intf_version = 1, +#endif +#ifdef HAVE_EXP_QP_BURST_CREATE_ENABLE_MULTI_PACKET_SEND_WR + /* Enable multi-packet send if supported. */ + .family_flags = + (priv->mps ? + IBV_EXP_QP_BURST_CREATE_ENABLE_MULTI_PACKET_SEND_WR : + 0), +#endif + }; + tmpl.if_qp = ibv_exp_query_intf(priv->ctx, &attr.params, &status); + if (tmpl.if_qp == NULL) { + ret = EINVAL; + ERROR("%p: QP interface family query failed with status %d", + (void *)dev, status); + goto error; + } + /* Clean up txq in case we're reinitializing it. */ + DEBUG("%p: cleaning-up old txq just in case", (void *)txq); + txq_cleanup(txq); + *txq = tmpl; + txq->poll_cnt = txq->if_cq->poll_cnt; +#if MLX5_PMD_MAX_INLINE > 0 + txq->send_pending_inline = txq->if_qp->send_pending_inline; +#ifdef HAVE_VERBS_VLAN_INSERTION + txq->send_pending_inline_vlan = txq->if_qp->send_pending_inline_vlan; +#endif +#endif +#if MLX5_PMD_SGE_WR_N > 1 + txq->send_pending_sg_list = txq->if_qp->send_pending_sg_list; +#ifdef HAVE_VERBS_VLAN_INSERTION + txq->send_pending_sg_list_vlan = txq->if_qp->send_pending_sg_list_vlan; +#endif +#endif + txq->send_pending = txq->if_qp->send_pending; +#ifdef HAVE_VERBS_VLAN_INSERTION + txq->send_pending_vlan = txq->if_qp->send_pending_vlan; +#endif + txq->send_flush = txq->if_qp->send_flush; + DEBUG("%p: txq updated with %p", (void *)txq, (void *)&tmpl); + /* Pre-register known mempools. */ + rte_mempool_walk(txq_mp2mr_iter, txq); + assert(ret == 0); + return 0; +error: + txq_cleanup(&tmpl); + assert(ret > 0); + return ret; +} + +/** + * DPDK callback to configure a TX queue. + * + * @param dev + * Pointer to Ethernet device structure. + * @param idx + * TX queue index. + * @param desc + * Number of descriptors to configure in queue. + * @param socket + * NUMA socket on which memory must be allocated. + * @param[in] conf + * Thresholds parameters. + * + * @return + * 0 on success, negative errno value on failure. + */ +int +mlx5_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, + unsigned int socket, const struct rte_eth_txconf *conf) +{ + struct priv *priv = dev->data->dev_private; + struct txq *txq = (*priv->txqs)[idx]; + int ret; + + if (mlx5_is_secondary()) + return -E_RTE_SECONDARY; + + priv_lock(priv); + DEBUG("%p: configuring queue %u for %u descriptors", + (void *)dev, idx, desc); + if (idx >= priv->txqs_n) { + ERROR("%p: queue index out of range (%u >= %u)", + (void *)dev, idx, priv->txqs_n); + priv_unlock(priv); + return -EOVERFLOW; + } + if (txq != NULL) { + DEBUG("%p: reusing already allocated queue index %u (%p)", + (void *)dev, idx, (void *)txq); + if (priv->started) { + priv_unlock(priv); + return -EEXIST; + } + (*priv->txqs)[idx] = NULL; + txq_cleanup(txq); + } else { + txq = rte_calloc_socket("TXQ", 1, sizeof(*txq), 0, socket); + if (txq == NULL) { + ERROR("%p: unable to allocate queue index %u", + (void *)dev, idx); + priv_unlock(priv); + return -ENOMEM; + } + } + ret = txq_setup(dev, txq, desc, socket, conf); + if (ret) + rte_free(txq); + else { + txq->stats.idx = idx; + DEBUG("%p: adding TX queue %p to list", + (void *)dev, (void *)txq); + (*priv->txqs)[idx] = txq; + /* Update send callback. */ + dev->tx_pkt_burst = mlx5_tx_burst; + } + priv_unlock(priv); + return -ret; +} + +/** + * DPDK callback to release a TX queue. + * + * @param dpdk_txq + * Generic TX queue pointer. + */ +void +mlx5_tx_queue_release(void *dpdk_txq) +{ + struct txq *txq = (struct txq *)dpdk_txq; + struct priv *priv; + unsigned int i; + + if (mlx5_is_secondary()) + return; + + if (txq == NULL) + return; + priv = txq->priv; + priv_lock(priv); + for (i = 0; (i != priv->txqs_n); ++i) + if ((*priv->txqs)[i] == txq) { + DEBUG("%p: removing TX queue %p from list", + (void *)priv->dev, (void *)txq); + (*priv->txqs)[i] = NULL; + break; + } + txq_cleanup(txq); + rte_free(txq); + priv_unlock(priv); +} + +/** + * DPDK callback for TX in secondary processes. + * + * This function configures all queues from primary process information + * if necessary before reverting to the normal TX burst callback. + * + * @param dpdk_txq + * Generic pointer to TX queue structure. + * @param[in] pkts + * Packets to transmit. + * @param pkts_n + * Number of packets in array. + * + * @return + * Number of packets successfully transmitted (<= pkts_n). + */ +uint16_t +mlx5_tx_burst_secondary_setup(void *dpdk_txq, struct rte_mbuf **pkts, + uint16_t pkts_n) +{ + struct txq *txq = dpdk_txq; + struct priv *priv = mlx5_secondary_data_setup(txq->priv); + struct priv *primary_priv; + unsigned int index; + + if (priv == NULL) + return 0; + primary_priv = + mlx5_secondary_data[priv->dev->data->port_id].primary_priv; + /* Look for queue index in both private structures. */ + for (index = 0; index != priv->txqs_n; ++index) + if (((*primary_priv->txqs)[index] == txq) || + ((*priv->txqs)[index] == txq)) + break; + if (index == priv->txqs_n) + return 0; + txq = (*priv->txqs)[index]; + return priv->dev->tx_pkt_burst(txq, pkts, pkts_n); +} -- cgit 1.2.3-korg