X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fmlx4%2Fmlx4_flow.c;h=78bd452bdd52382f97d28df34de9c17b736a7360;hb=6d13ea8e8e49;hp=41423cd8c40ac01cf9a2d396476a8f94453550e6;hpb=eacaac7bae3639b3aa798f77bf079f56b254c572;p=dpdk.git diff --git a/drivers/net/mlx4/mlx4_flow.c b/drivers/net/mlx4/mlx4_flow.c index 41423cd8c4..78bd452bdd 100644 --- a/drivers/net/mlx4/mlx4_flow.c +++ b/drivers/net/mlx4/mlx4_flow.c @@ -1,34 +1,6 @@ -/*- - * BSD LICENSE - * - * Copyright 2017 6WIND S.A. - * Copyright 2017 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. +/* SPDX-License-Identifier: BSD-3-Clause + * Copyright 2017 6WIND S.A. + * Copyright 2017 Mellanox Technologies, Ltd */ /** @@ -56,8 +28,7 @@ #include #include -#include -#include +#include #include #include #include @@ -65,6 +36,7 @@ /* PMD headers. */ #include "mlx4.h" +#include "mlx4_glue.h" #include "mlx4_flow.h" #include "mlx4_rxtx.h" #include "mlx4_utils.h" @@ -98,10 +70,103 @@ struct mlx4_flow_proc_item { struct mlx4_drop { struct ibv_qp *qp; /**< QP target. */ struct ibv_cq *cq; /**< CQ associated with above QP. */ - struct priv *priv; /**< Back pointer to private data. */ + struct mlx4_priv *priv; /**< Back pointer to private data. */ uint32_t refcnt; /**< Reference count. */ }; +/** + * Convert supported RSS hash field types between DPDK and Verbs formats. + * + * This function returns the supported (default) set when @p types has + * special value 0. + * + * @param priv + * Pointer to private structure. + * @param types + * Depending on @p verbs_to_dpdk, hash types in either DPDK (see struct + * rte_eth_rss_conf) or Verbs format. + * @param verbs_to_dpdk + * A zero value converts @p types from DPDK to Verbs, a nonzero value + * performs the reverse operation. + * + * @return + * Converted RSS hash fields on success, (uint64_t)-1 otherwise and + * rte_errno is set. + */ +uint64_t +mlx4_conv_rss_types(struct mlx4_priv *priv, uint64_t types, int verbs_to_dpdk) +{ + enum { + INNER, + IPV4, IPV4_1, IPV4_2, IPV6, IPV6_1, IPV6_2, IPV6_3, + TCP, UDP, + IPV4_TCP, IPV4_UDP, IPV6_TCP, IPV6_TCP_1, IPV6_UDP, IPV6_UDP_1, + }; + enum { + VERBS_IPV4 = IBV_RX_HASH_SRC_IPV4 | IBV_RX_HASH_DST_IPV4, + VERBS_IPV6 = IBV_RX_HASH_SRC_IPV6 | IBV_RX_HASH_DST_IPV6, + VERBS_TCP = IBV_RX_HASH_SRC_PORT_TCP | IBV_RX_HASH_DST_PORT_TCP, + VERBS_UDP = IBV_RX_HASH_SRC_PORT_UDP | IBV_RX_HASH_DST_PORT_UDP, + }; + static const uint64_t dpdk[] = { + [INNER] = 0, + [IPV4] = ETH_RSS_IPV4, + [IPV4_1] = ETH_RSS_FRAG_IPV4, + [IPV4_2] = ETH_RSS_NONFRAG_IPV4_OTHER, + [IPV6] = ETH_RSS_IPV6, + [IPV6_1] = ETH_RSS_FRAG_IPV6, + [IPV6_2] = ETH_RSS_NONFRAG_IPV6_OTHER, + [IPV6_3] = ETH_RSS_IPV6_EX, + [TCP] = 0, + [UDP] = 0, + [IPV4_TCP] = ETH_RSS_NONFRAG_IPV4_TCP, + [IPV4_UDP] = ETH_RSS_NONFRAG_IPV4_UDP, + [IPV6_TCP] = ETH_RSS_NONFRAG_IPV6_TCP, + [IPV6_TCP_1] = ETH_RSS_IPV6_TCP_EX, + [IPV6_UDP] = ETH_RSS_NONFRAG_IPV6_UDP, + [IPV6_UDP_1] = ETH_RSS_IPV6_UDP_EX, + }; + static const uint64_t verbs[RTE_DIM(dpdk)] = { + [INNER] = IBV_RX_HASH_INNER, + [IPV4] = VERBS_IPV4, + [IPV4_1] = VERBS_IPV4, + [IPV4_2] = VERBS_IPV4, + [IPV6] = VERBS_IPV6, + [IPV6_1] = VERBS_IPV6, + [IPV6_2] = VERBS_IPV6, + [IPV6_3] = VERBS_IPV6, + [TCP] = VERBS_TCP, + [UDP] = VERBS_UDP, + [IPV4_TCP] = VERBS_IPV4 | VERBS_TCP, + [IPV4_UDP] = VERBS_IPV4 | VERBS_UDP, + [IPV6_TCP] = VERBS_IPV6 | VERBS_TCP, + [IPV6_TCP_1] = VERBS_IPV6 | VERBS_TCP, + [IPV6_UDP] = VERBS_IPV6 | VERBS_UDP, + [IPV6_UDP_1] = VERBS_IPV6 | VERBS_UDP, + }; + const uint64_t *in = verbs_to_dpdk ? verbs : dpdk; + const uint64_t *out = verbs_to_dpdk ? dpdk : verbs; + uint64_t seen = 0; + uint64_t conv = 0; + unsigned int i; + + if (!types) { + if (!verbs_to_dpdk) + return priv->hw_rss_sup; + types = priv->hw_rss_sup; + } + for (i = 0; i != RTE_DIM(dpdk); ++i) + if (in[i] && (types & in[i]) == in[i]) { + seen |= types & in[i]; + conv |= out[i]; + } + if ((verbs_to_dpdk || (conv & priv->hw_rss_sup) == conv) && + !(types & ~seen)) + return conv; + rte_errno = ENOTSUP; + return (uint64_t)-1; +} + /** * Merge Ethernet pattern item into flow rule handle. * @@ -139,9 +204,7 @@ mlx4_flow_merge_eth(struct rte_flow *flow, const char *msg; unsigned int i; - if (!mask) { - flow->promisc = 1; - } else { + if (mask) { uint32_t sum_dst = 0; uint32_t sum_src = 0; @@ -183,6 +246,12 @@ mlx4_flow_merge_eth(struct rte_flow *flow, .type = IBV_FLOW_SPEC_ETH, .size = sizeof(*eth), }; + if (!mask) { + eth->val.dst_mac[0] = 0xff; + flow->ibv_attr->type = IBV_FLOW_ATTR_ALL_DEFAULT; + flow->promisc = 1; + return 0; + } memcpy(eth->val.dst_mac, spec->dst.addr_bytes, ETHER_ADDR_LEN); memcpy(eth->mask.dst_mac, mask->dst.addr_bytes, ETHER_ADDR_LEN); /* Remove unwanted bits from values. */ @@ -245,6 +314,8 @@ mlx4_flow_merge_vlan(struct rte_flow *flow, eth->val.vlan_tag = spec->tci; eth->mask.vlan_tag = mask->tci; eth->val.vlan_tag &= eth->mask.vlan_tag; + if (flow->ibv_attr->type == IBV_FLOW_ATTR_ALL_DEFAULT) + flow->ibv_attr->type = IBV_FLOW_ATTR_NORMAL; return 0; error: return rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM, @@ -321,6 +392,9 @@ error: * Additional mlx4-specific constraints on supported fields: * * - No support for partial masks. + * - Due to HW/FW limitation, flow rule priority is not taken into account + * when matching UDP destination ports, doing is therefore only supported + * at the highest priority level (0). * * @param[in, out] flow * Flow rule handle to update. @@ -346,12 +420,17 @@ mlx4_flow_merge_udp(struct rte_flow *flow, struct ibv_flow_spec_tcp_udp *udp; const char *msg; - if (!mask || + if (mask && ((uint16_t)(mask->hdr.src_port + 1) > UINT16_C(1) || (uint16_t)(mask->hdr.dst_port + 1) > UINT16_C(1))) { msg = "mlx4 does not support matching partial UDP fields"; goto error; } + if (mask && mask->hdr.dst_port && flow->priority) { + msg = "combining UDP destination port matching with a nonzero" + " priority level is not supported"; + goto error; + } if (!flow->ibv_attr) return 0; ++flow->ibv_attr->num_of_specs; @@ -406,7 +485,7 @@ mlx4_flow_merge_tcp(struct rte_flow *flow, struct ibv_flow_spec_tcp_udp *tcp; const char *msg; - if (!mask || + if (mask && ((uint16_t)(mask->hdr.src_port + 1) > UINT16_C(1) || (uint16_t)(mask->hdr.dst_port + 1) > UINT16_C(1))) { msg = "mlx4 does not support matching partial TCP fields"; @@ -583,7 +662,7 @@ static const struct mlx4_flow_proc_item mlx4_flow_proc_item_list[] = { * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -mlx4_flow_prepare(struct priv *priv, +mlx4_flow_prepare(struct mlx4_priv *priv, const struct rte_flow_attr *attr, const struct rte_flow_item pattern[], const struct rte_flow_action actions[], @@ -596,6 +675,7 @@ mlx4_flow_prepare(struct priv *priv, struct rte_flow temp = { .ibv_attr_size = sizeof(*temp.ibv_attr) }; struct rte_flow *flow = &temp; const char *msg = NULL; + int overlap; if (attr->group) return rte_flow_error_set @@ -610,12 +690,18 @@ mlx4_flow_prepare(struct priv *priv, return rte_flow_error_set (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ATTR_EGRESS, NULL, "egress is not supported"); + if (attr->transfer) + return rte_flow_error_set + (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ATTR_TRANSFER, + NULL, "transfer is not supported"); if (!attr->ingress) return rte_flow_error_set (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ATTR_INGRESS, NULL, "only ingress is supported"); fill: + overlap = 0; proc = mlx4_flow_proc_item_list; + flow->priority = attr->priority; /* Go over pattern. */ for (item = pattern; item->type; ++item) { const struct mlx4_flow_proc_item *next = NULL; @@ -661,32 +747,132 @@ fill: } /* Go over actions list. */ for (action = actions; action->type; ++action) { + /* This one may appear anywhere multiple times. */ + if (action->type == RTE_FLOW_ACTION_TYPE_VOID) + continue; + /* Fate-deciding actions may appear exactly once. */ + if (overlap) { + msg = "cannot combine several fate-deciding actions," + " choose between DROP, QUEUE or RSS"; + goto exit_action_not_supported; + } + overlap = 1; switch (action->type) { const struct rte_flow_action_queue *queue; + const struct rte_flow_action_rss *rss; + const uint8_t *rss_key; + uint32_t rss_key_len; + uint64_t fields; + unsigned int i; - case RTE_FLOW_ACTION_TYPE_VOID: - continue; case RTE_FLOW_ACTION_TYPE_DROP: flow->drop = 1; break; case RTE_FLOW_ACTION_TYPE_QUEUE: + if (flow->rss) + break; queue = action->conf; - if (queue->index >= priv->dev->data->nb_rx_queues) + if (queue->index >= ETH_DEV(priv)->data->nb_rx_queues) { + msg = "queue target index beyond number of" + " configured Rx queues"; goto exit_action_not_supported; - flow->queue = 1; - flow->queue_id = queue->index; + } + flow->rss = mlx4_rss_get + (priv, 0, mlx4_rss_hash_key_default, 1, + &queue->index); + if (!flow->rss) { + msg = "not enough resources for additional" + " single-queue RSS context"; + goto exit_action_not_supported; + } + break; + case RTE_FLOW_ACTION_TYPE_RSS: + if (flow->rss) + break; + rss = action->conf; + /* Default RSS configuration if none is provided. */ + if (rss->key_len) { + rss_key = rss->key; + rss_key_len = rss->key_len; + } else { + rss_key = mlx4_rss_hash_key_default; + rss_key_len = MLX4_RSS_HASH_KEY_SIZE; + } + /* Sanity checks. */ + for (i = 0; i < rss->queue_num; ++i) + if (rss->queue[i] >= + ETH_DEV(priv)->data->nb_rx_queues) + break; + if (i != rss->queue_num) { + msg = "queue index target beyond number of" + " configured Rx queues"; + goto exit_action_not_supported; + } + if (!rte_is_power_of_2(rss->queue_num)) { + msg = "for RSS, mlx4 requires the number of" + " queues to be a power of two"; + goto exit_action_not_supported; + } + if (rss_key_len != sizeof(flow->rss->key)) { + msg = "mlx4 supports exactly one RSS hash key" + " length: " + MLX4_STR_EXPAND(MLX4_RSS_HASH_KEY_SIZE); + goto exit_action_not_supported; + } + for (i = 1; i < rss->queue_num; ++i) + if (rss->queue[i] - rss->queue[i - 1] != 1) + break; + if (i != rss->queue_num) { + msg = "mlx4 requires RSS contexts to use" + " consecutive queue indices only"; + goto exit_action_not_supported; + } + if (rss->queue[0] % rss->queue_num) { + msg = "mlx4 requires the first queue of a RSS" + " context to be aligned on a multiple" + " of the context size"; + goto exit_action_not_supported; + } + if (rss->func && + rss->func != RTE_ETH_HASH_FUNCTION_TOEPLITZ) { + msg = "the only supported RSS hash function" + " is Toeplitz"; + goto exit_action_not_supported; + } + if (rss->level) { + msg = "a nonzero RSS encapsulation level is" + " not supported"; + goto exit_action_not_supported; + } + rte_errno = 0; + fields = mlx4_conv_rss_types(priv, rss->types, 0); + if (fields == (uint64_t)-1 && rte_errno) { + msg = "unsupported RSS hash type requested"; + goto exit_action_not_supported; + } + flow->rss = mlx4_rss_get + (priv, fields, rss_key, rss->queue_num, + rss->queue); + if (!flow->rss) { + msg = "either invalid parameters or not enough" + " resources for additional multi-queue" + " RSS context"; + goto exit_action_not_supported; + } break; default: goto exit_action_not_supported; } } - if (!flow->queue && !flow->drop) - return rte_flow_error_set - (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, - NULL, "no valid action"); + /* When fate is unknown, drop traffic. */ + if (!overlap) + flow->drop = 1; /* Validation ends here. */ - if (!addr) + if (!addr) { + if (flow->rss) + mlx4_rss_put(flow->rss); return 0; + } if (flow == &temp) { /* Allocate proper handle based on collected data. */ const struct mlx4_malloc_vec vec[] = { @@ -702,15 +888,19 @@ fill: }, }; - if (!mlx4_zmallocv(__func__, vec, RTE_DIM(vec))) + if (!mlx4_zmallocv(__func__, vec, RTE_DIM(vec))) { + if (temp.rss) + mlx4_rss_put(temp.rss); return rte_flow_error_set (error, -rte_errno, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, "flow rule handle allocation failure"); + } /* Most fields will be updated by second pass. */ *flow = (struct rte_flow){ .ibv_attr = temp.ibv_attr, .ibv_attr_size = sizeof(*flow->ibv_attr), + .rss = temp.rss, }; *flow->ibv_attr = (struct ibv_flow_attr){ .type = IBV_FLOW_ATTR_NORMAL, @@ -727,7 +917,7 @@ exit_item_not_supported: item, msg ? msg : "item not supported"); exit_action_not_supported: return rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ACTION, - action, "action not supported"); + action, msg ? msg : "action not supported"); } /** @@ -743,7 +933,7 @@ mlx4_flow_validate(struct rte_eth_dev *dev, const struct rte_flow_action actions[], struct rte_flow_error *error) { - struct priv *priv = dev->data->dev_private; + struct mlx4_priv *priv = dev->data->dev_private; return mlx4_flow_prepare(priv, attr, pattern, actions, error, NULL); } @@ -759,7 +949,7 @@ mlx4_flow_validate(struct rte_eth_dev *dev, * is set. */ static struct mlx4_drop * -mlx4_drop_get(struct priv *priv) +mlx4_drop_get(struct mlx4_priv *priv) { struct mlx4_drop *drop = priv->drop; @@ -776,24 +966,25 @@ mlx4_drop_get(struct priv *priv) .priv = priv, .refcnt = 1, }; - drop->cq = ibv_create_cq(priv->ctx, 1, NULL, NULL, 0); + drop->cq = mlx4_glue->create_cq(priv->ctx, 1, NULL, NULL, 0); if (!drop->cq) goto error; - drop->qp = ibv_create_qp(priv->pd, - &(struct ibv_qp_init_attr){ - .send_cq = drop->cq, - .recv_cq = drop->cq, - .qp_type = IBV_QPT_RAW_PACKET, - }); + drop->qp = mlx4_glue->create_qp + (priv->pd, + &(struct ibv_qp_init_attr){ + .send_cq = drop->cq, + .recv_cq = drop->cq, + .qp_type = IBV_QPT_RAW_PACKET, + }); if (!drop->qp) goto error; priv->drop = drop; return drop; error: if (drop->qp) - claim_zero(ibv_destroy_qp(drop->qp)); + claim_zero(mlx4_glue->destroy_qp(drop->qp)); if (drop->cq) - claim_zero(ibv_destroy_cq(drop->cq)); + claim_zero(mlx4_glue->destroy_cq(drop->cq)); if (drop) rte_free(drop); rte_errno = ENOMEM; @@ -813,8 +1004,8 @@ mlx4_drop_put(struct mlx4_drop *drop) if (--drop->refcnt) return; drop->priv->drop = NULL; - claim_zero(ibv_destroy_qp(drop->qp)); - claim_zero(ibv_destroy_cq(drop->cq)); + claim_zero(mlx4_glue->destroy_qp(drop->qp)); + claim_zero(mlx4_glue->destroy_cq(drop->cq)); rte_free(drop); } @@ -834,7 +1025,7 @@ mlx4_drop_put(struct mlx4_drop *drop) * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -mlx4_flow_toggle(struct priv *priv, +mlx4_flow_toggle(struct mlx4_priv *priv, struct rte_flow *flow, int enable, struct rte_flow_error *error) @@ -846,10 +1037,12 @@ mlx4_flow_toggle(struct priv *priv, if (!enable) { if (!flow->ibv_flow) return 0; - claim_zero(ibv_destroy_flow(flow->ibv_flow)); + claim_zero(mlx4_glue->destroy_flow(flow->ibv_flow)); flow->ibv_flow = NULL; if (flow->drop) mlx4_drop_put(priv->drop); + else if (flow->rss) + mlx4_rss_detach(flow->rss); return 0; } assert(flow->ibv_attr); @@ -857,10 +1050,12 @@ mlx4_flow_toggle(struct priv *priv, !priv->isolated && flow->ibv_attr->priority == MLX4_FLOW_PRIORITY_LAST) { if (flow->ibv_flow) { - claim_zero(ibv_destroy_flow(flow->ibv_flow)); + claim_zero(mlx4_glue->destroy_flow(flow->ibv_flow)); flow->ibv_flow = NULL; if (flow->drop) mlx4_drop_put(priv->drop); + else if (flow->rss) + mlx4_rss_detach(flow->rss); } err = EACCES; msg = ("priority level " @@ -868,26 +1063,46 @@ mlx4_flow_toggle(struct priv *priv, " is reserved when not in isolated mode"); goto error; } - if (flow->queue) { - struct rxq *rxq = NULL; + if (flow->rss) { + struct mlx4_rss *rss = flow->rss; + int missing = 0; + unsigned int i; - if (flow->queue_id < priv->dev->data->nb_rx_queues) - rxq = priv->dev->data->rx_queues[flow->queue_id]; + /* Stop at the first nonexistent target queue. */ + for (i = 0; i != rss->queues; ++i) + if (rss->queue_id[i] >= + ETH_DEV(priv)->data->nb_rx_queues || + !ETH_DEV(priv)->data->rx_queues[rss->queue_id[i]]) { + missing = 1; + break; + } if (flow->ibv_flow) { - if (!rxq ^ !flow->drop) + if (missing ^ !flow->drop) return 0; /* Verbs flow needs updating. */ - claim_zero(ibv_destroy_flow(flow->ibv_flow)); + claim_zero(mlx4_glue->destroy_flow(flow->ibv_flow)); flow->ibv_flow = NULL; if (flow->drop) mlx4_drop_put(priv->drop); + else + mlx4_rss_detach(rss); + } + if (!missing) { + err = mlx4_rss_attach(rss); + if (err) { + err = -err; + msg = "cannot create indirection table or hash" + " QP to associate flow rule with"; + goto error; + } + qp = rss->qp; } - if (rxq) - qp = rxq->qp; /* A missing target queue drops traffic implicitly. */ - flow->drop = !rxq; + flow->drop = missing; } if (flow->drop) { + if (flow->ibv_flow) + return 0; mlx4_drop_get(priv); if (!priv->drop) { err = rte_errno; @@ -899,11 +1114,13 @@ mlx4_flow_toggle(struct priv *priv, assert(qp); if (flow->ibv_flow) return 0; - flow->ibv_flow = ibv_create_flow(qp, flow->ibv_attr); + flow->ibv_flow = mlx4_glue->create_flow(qp, flow->ibv_attr); if (flow->ibv_flow) return 0; if (flow->drop) mlx4_drop_put(priv->drop); + else if (flow->rss) + mlx4_rss_detach(flow->rss); err = errno; msg = "flow rule rejected by device"; error: @@ -924,7 +1141,7 @@ mlx4_flow_create(struct rte_eth_dev *dev, const struct rte_flow_action actions[], struct rte_flow_error *error) { - struct priv *priv = dev->data->dev_private; + struct mlx4_priv *priv = dev->data->dev_private; struct rte_flow *flow; int err; @@ -946,6 +1163,8 @@ mlx4_flow_create(struct rte_eth_dev *dev, } return flow; } + if (flow->rss) + mlx4_rss_put(flow->rss); rte_flow_error_set(error, -err, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, error->message); rte_free(flow); @@ -963,7 +1182,7 @@ mlx4_flow_isolate(struct rte_eth_dev *dev, int enable, struct rte_flow_error *error) { - struct priv *priv = dev->data->dev_private; + struct mlx4_priv *priv = dev->data->dev_private; if (!!enable == !!priv->isolated) return 0; @@ -986,12 +1205,14 @@ mlx4_flow_destroy(struct rte_eth_dev *dev, struct rte_flow *flow, struct rte_flow_error *error) { - struct priv *priv = dev->data->dev_private; + struct mlx4_priv *priv = dev->data->dev_private; int err = mlx4_flow_toggle(priv, flow, 0, error); if (err) return err; LIST_REMOVE(flow, next); + if (flow->rss) + mlx4_rss_put(flow->rss); rte_free(flow); return 0; } @@ -1008,7 +1229,7 @@ static int mlx4_flow_flush(struct rte_eth_dev *dev, struct rte_flow_error *error) { - struct priv *priv = dev->data->dev_private; + struct mlx4_priv *priv = dev->data->dev_private; struct rte_flow *flow = LIST_FIRST(&priv->flows); while (flow) { @@ -1033,10 +1254,10 @@ mlx4_flow_flush(struct rte_eth_dev *dev, * Next configured VLAN ID or a high value (>= 4096) if there is none. */ static uint16_t -mlx4_flow_internal_next_vlan(struct priv *priv, uint16_t vlan) +mlx4_flow_internal_next_vlan(struct mlx4_priv *priv, uint16_t vlan) { while (vlan < 4096) { - if (priv->dev->data->vlan_filter_conf.ids[vlan / 64] & + if (ETH_DEV(priv)->data->vlan_filter_conf.ids[vlan / 64] & (UINT64_C(1) << (vlan % 64))) return vlan; ++vlan; @@ -1049,16 +1270,19 @@ mlx4_flow_internal_next_vlan(struct priv *priv, uint16_t vlan) * * Various flow rules are created depending on the mode the device is in: * - * 1. Promiscuous: port MAC + catch-all (VLAN filtering is ignored). - * 2. All multicast: port MAC/VLAN + catch-all multicast. - * 3. Otherwise: port MAC/VLAN + broadcast MAC/VLAN. + * 1. Promiscuous: + * port MAC + broadcast + catch-all (VLAN filtering is ignored). + * 2. All multicast: + * port MAC/VLAN + broadcast + catch-all multicast. + * 3. Otherwise: + * port MAC/VLAN + broadcast MAC/VLAN. * * About MAC flow rules: * * - MAC flow rules are generated from @p dev->data->mac_addrs * (@p priv->mac array). * - An additional flow rule for Ethernet broadcasts is also generated. - * - All these are per-VLAN if @p dev->data->dev_conf.rxmode.hw_vlan_filter + * - All these are per-VLAN if @p DEV_RX_OFFLOAD_VLAN_FILTER * is enabled and VLAN filters are configured. * * @param priv @@ -1070,7 +1294,7 @@ mlx4_flow_internal_next_vlan(struct priv *priv, uint16_t vlan) * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -mlx4_flow_internal(struct priv *priv, struct rte_flow_error *error) +mlx4_flow_internal(struct mlx4_priv *priv, struct rte_flow_error *error) { struct rte_flow_attr attr = { .priority = MLX4_FLOW_PRIORITY_LAST, @@ -1104,31 +1328,50 @@ mlx4_flow_internal(struct priv *priv, struct rte_flow_error *error) .type = RTE_FLOW_ITEM_TYPE_END, }, }; + /* + * Round number of queues down to their previous power of 2 to + * comply with RSS context limitations. Extra queues silently do not + * get RSS by default. + */ + uint32_t queues = + rte_align32pow2(ETH_DEV(priv)->data->nb_rx_queues + 1) >> 1; + uint16_t queue[queues]; + struct rte_flow_action_rss action_rss = { + .func = RTE_ETH_HASH_FUNCTION_DEFAULT, + .level = 0, + .types = 0, + .key_len = MLX4_RSS_HASH_KEY_SIZE, + .queue_num = queues, + .key = mlx4_rss_hash_key_default, + .queue = queue, + }; struct rte_flow_action actions[] = { { - .type = RTE_FLOW_ACTION_TYPE_QUEUE, - .conf = &(struct rte_flow_action_queue){ - .index = 0, - }, + .type = RTE_FLOW_ACTION_TYPE_RSS, + .conf = &action_rss, }, { .type = RTE_FLOW_ACTION_TYPE_END, }, }; - struct ether_addr *rule_mac = ð_spec.dst; + struct rte_ether_addr *rule_mac = ð_spec.dst; rte_be16_t *rule_vlan = - priv->dev->data->dev_conf.rxmode.hw_vlan_filter && - !priv->dev->data->promiscuous ? + (ETH_DEV(priv)->data->dev_conf.rxmode.offloads & + DEV_RX_OFFLOAD_VLAN_FILTER) && + !ETH_DEV(priv)->data->promiscuous ? &vlan_spec.tci : NULL; - int broadcast = - !priv->dev->data->promiscuous && - !priv->dev->data->all_multicast; uint16_t vlan = 0; struct rte_flow *flow; unsigned int i; int err = 0; + /* Nothing to be done if there are no Rx queues. */ + if (!queues) + goto error; + /* Prepare default RSS configuration. */ + for (i = 0; i != queues; ++i) + queue[i] = i; /* * Set up VLAN item if filtering is enabled and at least one VLAN * filter is configured. @@ -1147,8 +1390,8 @@ next_vlan: rule_vlan = NULL; } } - for (i = 0; i != RTE_DIM(priv->mac) + broadcast; ++i) { - const struct ether_addr *mac; + for (i = 0; i != RTE_DIM(priv->mac) + 1; ++i) { + const struct rte_ether_addr *mac; /* Broadcasts are handled by an extra iteration. */ if (i < RTE_DIM(priv->mac)) @@ -1171,6 +1414,7 @@ next_vlan: assert(flow->ibv_attr->type == IBV_FLOW_ATTR_NORMAL); assert(flow->ibv_attr->num_of_specs == 1); assert(eth->type == IBV_FLOW_SPEC_ETH); + assert(flow->rss); if (rule_vlan && (eth->val.vlan_tag != *rule_vlan || eth->mask.vlan_tag != RTE_BE16(0x0fff))) @@ -1183,13 +1427,18 @@ next_vlan: eth->val.src_mac[j] != UINT8_C(0x00) || eth->mask.src_mac[j] != UINT8_C(0x00)) break; - if (j == sizeof(mac->addr_bytes)) - break; + if (j != sizeof(mac->addr_bytes)) + continue; + if (flow->rss->queues != queues || + memcmp(flow->rss->queue_id, action_rss.queue, + queues * sizeof(flow->rss->queue_id[0]))) + continue; + break; } if (!flow || !flow->internal) { /* Not found, create a new flow rule. */ memcpy(rule_mac, mac, sizeof(*mac)); - flow = mlx4_flow_create(priv->dev, &attr, pattern, + flow = mlx4_flow_create(ETH_DEV(priv), &attr, pattern, actions, error); if (!flow) { err = -rte_errno; @@ -1205,31 +1454,39 @@ next_vlan: goto next_vlan; } /* Take care of promiscuous and all multicast flow rules. */ - if (!broadcast) { + if (ETH_DEV(priv)->data->promiscuous || + ETH_DEV(priv)->data->all_multicast) { for (flow = LIST_FIRST(&priv->flows); flow && flow->internal; flow = LIST_NEXT(flow, next)) { - if (priv->dev->data->promiscuous) { + if (ETH_DEV(priv)->data->promiscuous) { if (flow->promisc) break; } else { - assert(priv->dev->data->all_multicast); + assert(ETH_DEV(priv)->data->all_multicast); if (flow->allmulti) break; } } + if (flow && flow->internal) { + assert(flow->rss); + if (flow->rss->queues != queues || + memcmp(flow->rss->queue_id, action_rss.queue, + queues * sizeof(flow->rss->queue_id[0]))) + flow = NULL; + } if (!flow || !flow->internal) { /* Not found, create a new flow rule. */ - if (priv->dev->data->promiscuous) { + if (ETH_DEV(priv)->data->promiscuous) { pattern[1].spec = NULL; pattern[1].mask = NULL; } else { - assert(priv->dev->data->all_multicast); + assert(ETH_DEV(priv)->data->all_multicast); pattern[1].spec = ð_allmulti; pattern[1].mask = ð_allmulti; } pattern[2] = pattern[3]; - flow = mlx4_flow_create(priv->dev, &attr, pattern, + flow = mlx4_flow_create(ETH_DEV(priv), &attr, pattern, actions, error); if (!flow) { err = -rte_errno; @@ -1246,7 +1503,8 @@ error: struct rte_flow *next = LIST_NEXT(flow, next); if (!flow->select) - claim_zero(mlx4_flow_destroy(priv->dev, flow, error)); + claim_zero(mlx4_flow_destroy(ETH_DEV(priv), flow, + error)); else flow->select = 0; flow = next; @@ -1270,7 +1528,7 @@ error: * 0 on success, a negative errno value otherwise and rte_errno is set. */ int -mlx4_flow_sync(struct priv *priv, struct rte_flow_error *error) +mlx4_flow_sync(struct mlx4_priv *priv, struct rte_flow_error *error) { struct rte_flow *flow; int ret; @@ -1284,7 +1542,8 @@ mlx4_flow_sync(struct priv *priv, struct rte_flow_error *error) for (flow = LIST_FIRST(&priv->flows); flow && flow->internal; flow = LIST_FIRST(&priv->flows)) - claim_zero(mlx4_flow_destroy(priv->dev, flow, error)); + claim_zero(mlx4_flow_destroy(ETH_DEV(priv), flow, + error)); } else { /* Refresh internal rules. */ ret = mlx4_flow_internal(priv, error); @@ -1292,9 +1551,7 @@ mlx4_flow_sync(struct priv *priv, struct rte_flow_error *error) return ret; } /* Toggle the remaining flow rules . */ - for (flow = LIST_FIRST(&priv->flows); - flow; - flow = LIST_NEXT(flow, next)) { + LIST_FOREACH(flow, &priv->flows, next) { ret = mlx4_flow_toggle(priv, flow, priv->started, error); if (ret) return ret; @@ -1314,12 +1571,13 @@ mlx4_flow_sync(struct priv *priv, struct rte_flow_error *error) * Pointer to private structure. */ void -mlx4_flow_clean(struct priv *priv) +mlx4_flow_clean(struct mlx4_priv *priv) { struct rte_flow *flow; while ((flow = LIST_FIRST(&priv->flows))) - mlx4_flow_destroy(priv->dev, flow, NULL); + mlx4_flow_destroy(ETH_DEV(priv), flow, NULL); + assert(LIST_EMPTY(&priv->rss)); } static const struct rte_flow_ops mlx4_flow_ops = {