X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fmlx5%2Fmlx5_flow_verbs.c;h=29cd694752f09c570b7a9b6183715dc87ac48a58;hb=693c7d4b1e12024cd429ef563e9df5cc2c477aee;hp=a435d65082cbf0fda774a75a7f177a5bfe099ea6;hpb=246636411536a165637fc8e970192cc5b62de0f5;p=dpdk.git diff --git a/drivers/net/mlx5/mlx5_flow_verbs.c b/drivers/net/mlx5/mlx5_flow_verbs.c index a435d65082..29cd694752 100644 --- a/drivers/net/mlx5/mlx5_flow_verbs.c +++ b/drivers/net/mlx5/mlx5_flow_verbs.c @@ -8,30 +8,130 @@ #include #include -/* Verbs header. */ -/* ISO C doesn't support unnamed structs/unions, disabling -pedantic. */ -#ifdef PEDANTIC -#pragma GCC diagnostic ignored "-Wpedantic" -#endif -#include -#ifdef PEDANTIC -#pragma GCC diagnostic error "-Wpedantic" -#endif - #include #include -#include -#include +#include #include #include #include #include -#include "mlx5.h" +#include +#include +#include + #include "mlx5_defs.h" -#include "mlx5_prm.h" -#include "mlx5_glue.h" +#include "mlx5.h" #include "mlx5_flow.h" +#include "mlx5_rx.h" + +#define VERBS_SPEC_INNER(item_flags) \ + (!!((item_flags) & MLX5_FLOW_LAYER_TUNNEL) ? IBV_FLOW_SPEC_INNER : 0) + +/* Verbs specification header. */ +struct ibv_spec_header { + enum ibv_flow_spec_type type; + uint16_t size; +}; + +/** + * Discover the maximum number of priority available. + * + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] vprio + * Expected result variants. + * @param[in] vprio_n + * Number of entries in @p vprio array. + * @return + * Number of supported flow priority on success, a negative errno + * value otherwise and rte_errno is set. + */ +static int +flow_verbs_discover_priorities(struct rte_eth_dev *dev, + const uint16_t *vprio, int vprio_n) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct { + struct ibv_flow_attr attr; + struct ibv_flow_spec_eth eth; + struct ibv_flow_spec_action_drop drop; + } flow_attr = { + .attr = { + .num_of_specs = 2, + .port = (uint8_t)priv->dev_port, + }, + .eth = { + .type = IBV_FLOW_SPEC_ETH, + .size = sizeof(struct ibv_flow_spec_eth), + }, + .drop = { + .size = sizeof(struct ibv_flow_spec_action_drop), + .type = IBV_FLOW_SPEC_ACTION_DROP, + }, + }; + struct ibv_flow *flow; + struct mlx5_hrxq *drop = priv->drop_queue.hrxq; + int i; + int priority = 0; + +#if defined(HAVE_MLX5DV_DR_DEVX_PORT) || defined(HAVE_MLX5DV_DR_DEVX_PORT_V35) + /* If DevX supported, driver must support 16 verbs flow priorities. */ + priority = 16; + goto out; +#endif + if (!drop->qp) { + rte_errno = ENOTSUP; + return -rte_errno; + } + for (i = 0; i != vprio_n; i++) { + flow_attr.attr.priority = vprio[i] - 1; + flow = mlx5_glue->create_flow(drop->qp, &flow_attr.attr); + if (!flow) + break; + claim_zero(mlx5_glue->destroy_flow(flow)); + priority = vprio[i]; + } +#if defined(HAVE_MLX5DV_DR_DEVX_PORT) || defined(HAVE_MLX5DV_DR_DEVX_PORT_V35) +out: +#endif + DRV_LOG(INFO, "port %u supported flow priorities:" + " 0-%d for ingress or egress root table," + " 0-%d for non-root table or transfer root table.", + dev->data->port_id, priority - 2, + MLX5_NON_ROOT_FLOW_MAX_PRIO - 1); + return priority; +} + +/** + * Get Verbs flow counter by index. + * + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] idx + * mlx5 flow counter index in the container. + * @param[out] ppool + * mlx5 flow counter pool in the container, + * + * @return + * A pointer to the counter, NULL otherwise. + */ +static struct mlx5_flow_counter * +flow_verbs_counter_get_by_idx(struct rte_eth_dev *dev, + uint32_t idx, + struct mlx5_flow_counter_pool **ppool) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_counter_mng *cmng = &priv->sh->cmng; + struct mlx5_flow_counter_pool *pool; + + idx = (idx - 1) & (MLX5_CNT_SHARED_OFFSET - 1); + pool = cmng->pools[idx / MLX5_COUNTERS_PER_POOL]; + MLX5_ASSERT(pool); + if (ppool) + *ppool = pool; + return MLX5_POOL_GET_CNT(pool, idx % MLX5_COUNTERS_PER_POOL); +} /** * Create Verbs flow counter with Verbs library. @@ -52,39 +152,42 @@ flow_verbs_counter_create(struct rte_eth_dev *dev, struct mlx5_flow_counter *counter) { #if defined(HAVE_IBV_DEVICE_COUNTERS_SET_V42) - struct priv *priv = dev->data->dev_private; + struct mlx5_priv *priv = dev->data->dev_private; + struct ibv_context *ctx = priv->sh->cdev->ctx; struct ibv_counter_set_init_attr init = { - .counter_set_id = counter->id}; + .counter_set_id = counter->shared_info.id}; - counter->cs = mlx5_glue->create_counter_set(priv->ctx, &init); - if (!counter->cs) { + counter->dcs_when_free = mlx5_glue->create_counter_set(ctx, &init); + if (!counter->dcs_when_free) { rte_errno = ENOTSUP; return -ENOTSUP; } return 0; #elif defined(HAVE_IBV_DEVICE_COUNTERS_SET_V45) - struct priv *priv = dev->data->dev_private; + struct mlx5_priv *priv = dev->data->dev_private; + struct ibv_context *ctx = priv->sh->cdev->ctx; struct ibv_counters_init_attr init = {0}; - struct ibv_counter_attach_attr attach = {0}; + struct ibv_counter_attach_attr attach; int ret; - counter->cs = mlx5_glue->create_counters(priv->ctx, &init); - if (!counter->cs) { + memset(&attach, 0, sizeof(attach)); + counter->dcs_when_free = mlx5_glue->create_counters(ctx, &init); + if (!counter->dcs_when_free) { rte_errno = ENOTSUP; return -ENOTSUP; } attach.counter_desc = IBV_COUNTER_PACKETS; attach.index = 0; - ret = mlx5_glue->attach_counters(counter->cs, &attach, NULL); + ret = mlx5_glue->attach_counters(counter->dcs_when_free, &attach, NULL); if (!ret) { attach.counter_desc = IBV_COUNTER_BYTES; attach.index = 1; ret = mlx5_glue->attach_counters - (counter->cs, &attach, NULL); + (counter->dcs_when_free, &attach, NULL); } if (ret) { - claim_zero(mlx5_glue->destroy_counters(counter->cs)); - counter->cs = NULL; + claim_zero(mlx5_glue->destroy_counters(counter->dcs_when_free)); + counter->dcs_when_free = NULL; rte_errno = ret; return -ret; } @@ -102,69 +205,107 @@ flow_verbs_counter_create(struct rte_eth_dev *dev, * * @param[in] dev * Pointer to the Ethernet device structure. - * @param[in] shared - * Indicate if this counter is shared with other flows. * @param[in] id * Counter identifier. * * @return - * A pointer to the counter, NULL otherwise and rte_errno is set. + * Index to the counter, 0 otherwise and rte_errno is set. */ -static struct mlx5_flow_counter * -flow_verbs_counter_new(struct rte_eth_dev *dev, uint32_t shared, uint32_t id) +static uint32_t +flow_verbs_counter_new(struct rte_eth_dev *dev, uint32_t id __rte_unused) { - struct priv *priv = dev->data->dev_private; - struct mlx5_flow_counter *cnt; + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_counter_mng *cmng = &priv->sh->cmng; + struct mlx5_flow_counter_pool *pool = NULL; + struct mlx5_flow_counter *cnt = NULL; + uint32_t n_valid = cmng->n_valid; + uint32_t pool_idx, cnt_idx; + uint32_t i; int ret; - LIST_FOREACH(cnt, &priv->flow_counters, next) { - if (!cnt->shared || cnt->shared != shared) - continue; - if (cnt->id != id) + for (pool_idx = 0; pool_idx < n_valid; ++pool_idx) { + pool = cmng->pools[pool_idx]; + if (!pool) continue; - cnt->ref_cnt++; - return cnt; + cnt = TAILQ_FIRST(&pool->counters[0]); + if (cnt) + break; } - cnt = rte_calloc(__func__, 1, sizeof(*cnt), 0); if (!cnt) { - rte_errno = ENOMEM; - return NULL; + struct mlx5_flow_counter_pool **pools; + uint32_t size; + + if (n_valid == cmng->n) { + /* Resize the container pool array. */ + size = sizeof(struct mlx5_flow_counter_pool *) * + (n_valid + MLX5_CNT_CONTAINER_RESIZE); + pools = mlx5_malloc(MLX5_MEM_ZERO, size, 0, + SOCKET_ID_ANY); + if (!pools) + return 0; + if (n_valid) { + memcpy(pools, cmng->pools, + sizeof(struct mlx5_flow_counter_pool *) * + n_valid); + mlx5_free(cmng->pools); + } + cmng->pools = pools; + cmng->n += MLX5_CNT_CONTAINER_RESIZE; + } + /* Allocate memory for new pool*/ + size = sizeof(*pool) + sizeof(*cnt) * MLX5_COUNTERS_PER_POOL; + pool = mlx5_malloc(MLX5_MEM_ZERO, size, 0, SOCKET_ID_ANY); + if (!pool) + return 0; + for (i = 0; i < MLX5_COUNTERS_PER_POOL; ++i) { + cnt = MLX5_POOL_GET_CNT(pool, i); + TAILQ_INSERT_HEAD(&pool->counters[0], cnt, next); + } + cnt = MLX5_POOL_GET_CNT(pool, 0); + cmng->pools[n_valid] = pool; + pool_idx = n_valid; + cmng->n_valid++; } - cnt->id = id; - cnt->shared = shared; - cnt->ref_cnt = 1; - cnt->hits = 0; - cnt->bytes = 0; + TAILQ_REMOVE(&pool->counters[0], cnt, next); + i = MLX5_CNT_ARRAY_IDX(pool, cnt); + cnt_idx = MLX5_MAKE_CNT_IDX(pool_idx, i); /* Create counter with Verbs. */ ret = flow_verbs_counter_create(dev, cnt); if (!ret) { - LIST_INSERT_HEAD(&priv->flow_counters, cnt, next); - return cnt; + cnt->dcs_when_active = cnt->dcs_when_free; + cnt->hits = 0; + cnt->bytes = 0; + return cnt_idx; } + TAILQ_INSERT_HEAD(&pool->counters[0], cnt, next); /* Some error occurred in Verbs library. */ - rte_free(cnt); rte_errno = -ret; - return NULL; + return 0; } /** * Release a flow counter. * + * @param[in] dev + * Pointer to the Ethernet device structure. * @param[in] counter - * Pointer to the counter handler. + * Index to the counter handler. */ static void -flow_verbs_counter_release(struct mlx5_flow_counter *counter) +flow_verbs_counter_release(struct rte_eth_dev *dev, uint32_t counter) { - if (--counter->ref_cnt == 0) { + struct mlx5_flow_counter_pool *pool; + struct mlx5_flow_counter *cnt; + + cnt = flow_verbs_counter_get_by_idx(dev, counter, &pool); #if defined(HAVE_IBV_DEVICE_COUNTERS_SET_V42) - claim_zero(mlx5_glue->destroy_counter_set(counter->cs)); + claim_zero(mlx5_glue->destroy_counter_set + ((struct ibv_counter_set *)cnt->dcs_when_active)); #elif defined(HAVE_IBV_DEVICE_COUNTERS_SET_V45) - claim_zero(mlx5_glue->destroy_counters(counter->cs)); + claim_zero(mlx5_glue->destroy_counters + ((struct ibv_counters *)cnt->dcs_when_active)); #endif - LIST_REMOVE(counter, next); - rte_free(counter); - } + TAILQ_INSERT_HEAD(&pool->counters[0], cnt, next); } /** @@ -180,12 +321,16 @@ flow_verbs_counter_query(struct rte_eth_dev *dev __rte_unused, { #if defined(HAVE_IBV_DEVICE_COUNTERS_SET_V42) || \ defined(HAVE_IBV_DEVICE_COUNTERS_SET_V45) - if (flow->actions & MLX5_FLOW_ACTION_COUNT) { + if (flow->counter) { + struct mlx5_flow_counter_pool *pool; + struct mlx5_flow_counter *cnt = flow_verbs_counter_get_by_idx + (dev, flow->counter, &pool); struct rte_flow_query_count *qc = data; uint64_t counters[2] = {0, 0}; #if defined(HAVE_IBV_DEVICE_COUNTERS_SET_V42) struct ibv_query_counter_set_attr query_cs_attr = { - .cs = flow->counter->cs, + .dcs_when_free = (struct ibv_counter_set *) + cnt->dcs_when_active, .query_flags = IBV_COUNTER_SET_FORCE_UPDATE, }; struct ibv_counter_set_data query_out = { @@ -196,7 +341,7 @@ flow_verbs_counter_query(struct rte_eth_dev *dev __rte_unused, &query_out); #elif defined(HAVE_IBV_DEVICE_COUNTERS_SET_V45) int err = mlx5_glue->query_counters - (flow->counter->cs, counters, + ((struct ibv_counters *)cnt->dcs_when_active, counters, RTE_DIM(counters), IBV_READ_COUNTERS_ATTR_PREFER_CACHED); #endif @@ -208,11 +353,11 @@ flow_verbs_counter_query(struct rte_eth_dev *dev __rte_unused, "cannot read counter"); qc->hits_set = 1; qc->bytes_set = 1; - qc->hits = counters[0] - flow->counter->hits; - qc->bytes = counters[1] - flow->counter->bytes; + qc->hits = counters[0] - cnt->hits; + qc->bytes = counters[1] - cnt->bytes; if (qc->reset) { - flow->counter->hits = counters[0]; - flow->counter->bytes = counters[1]; + cnt->hits = counters[0]; + cnt->bytes = counters[1]; } return 0; } @@ -231,27 +376,27 @@ flow_verbs_counter_query(struct rte_eth_dev *dev __rte_unused, } /** - * Add a verbs item specification into @p flow. + * Add a verbs item specification into @p verbs. * - * @param[in, out] flow - * Pointer to flow structure. + * @param[out] verbs + * Pointer to verbs structure. * @param[in] src * Create specification. * @param[in] size * Size in bytes of the specification to copy. */ static void -flow_verbs_spec_add(struct mlx5_flow *flow, void *src, unsigned int size) +flow_verbs_spec_add(struct mlx5_flow_verbs_workspace *verbs, + void *src, unsigned int size) { - struct mlx5_flow_verbs *verbs = &flow->verbs; + void *dst; - if (verbs->specs) { - void *dst; - - dst = (void *)(verbs->specs + verbs->size); - memcpy(dst, src, size); - ++verbs->attr->num_of_specs; - } + if (!verbs) + return; + MLX5_ASSERT(verbs->specs); + dst = (void *)(verbs->specs + verbs->size); + memcpy(dst, src, size); + ++verbs->attr.num_of_specs; verbs->size += size; } @@ -260,24 +405,23 @@ flow_verbs_spec_add(struct mlx5_flow *flow, void *src, unsigned int size) * the input is valid and that there is space to insert the requested item * into the flow. * + * @param[in, out] dev_flow + * Pointer to dev_flow structure. * @param[in] item * Item specification. * @param[in] item_flags - * Bit field with all detected items. - * @param[in, out] dev_flow - * Pointer to dev_flow structure. + * Parsed item flags. */ static void -flow_verbs_translate_item_eth(const struct rte_flow_item *item, - uint64_t *item_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_item_eth(struct mlx5_flow *dev_flow, + const struct rte_flow_item *item, + uint64_t item_flags) { const struct rte_flow_item_eth *spec = item->spec; const struct rte_flow_item_eth *mask = item->mask; - const int tunnel = !!(*item_flags & MLX5_FLOW_LAYER_TUNNEL); const unsigned int size = sizeof(struct ibv_flow_spec_eth); struct ibv_flow_spec_eth eth = { - .type = IBV_FLOW_SPEC_ETH | (tunnel ? IBV_FLOW_SPEC_INNER : 0), + .type = IBV_FLOW_SPEC_ETH | VERBS_SPEC_INNER(item_flags), .size = size, }; @@ -286,23 +430,24 @@ flow_verbs_translate_item_eth(const struct rte_flow_item *item, if (spec) { unsigned int i; - memcpy(ð.val.dst_mac, spec->dst.addr_bytes, ETHER_ADDR_LEN); - memcpy(ð.val.src_mac, spec->src.addr_bytes, ETHER_ADDR_LEN); + memcpy(ð.val.dst_mac, spec->dst.addr_bytes, + RTE_ETHER_ADDR_LEN); + memcpy(ð.val.src_mac, spec->src.addr_bytes, + RTE_ETHER_ADDR_LEN); eth.val.ether_type = spec->type; - memcpy(ð.mask.dst_mac, mask->dst.addr_bytes, ETHER_ADDR_LEN); - memcpy(ð.mask.src_mac, mask->src.addr_bytes, ETHER_ADDR_LEN); + memcpy(ð.mask.dst_mac, mask->dst.addr_bytes, + RTE_ETHER_ADDR_LEN); + memcpy(ð.mask.src_mac, mask->src.addr_bytes, + RTE_ETHER_ADDR_LEN); eth.mask.ether_type = mask->type; /* Remove unwanted bits from values. */ - for (i = 0; i < ETHER_ADDR_LEN; ++i) { + for (i = 0; i < RTE_ETHER_ADDR_LEN; ++i) { eth.val.dst_mac[i] &= eth.mask.dst_mac[i]; eth.val.src_mac[i] &= eth.mask.src_mac[i]; } eth.val.ether_type &= eth.mask.ether_type; - dev_flow->verbs.attr->priority = MLX5_PRIORITY_MAP_L2; } - flow_verbs_spec_add(dev_flow, ð, size); - *item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L2 : - MLX5_FLOW_LAYER_OUTER_L2; + flow_verbs_spec_add(&dev_flow->verbs, ð, size); } /** @@ -344,24 +489,24 @@ flow_verbs_item_vlan_update(struct ibv_flow_attr *attr, * the input is valid and that there is space to insert the requested item * into the flow. * - * @param[in] item - * Item specification. - * @param[in, out] item_flags - * Bit mask that holds all detected items. * @param[in, out] dev_flow * Pointer to dev_flow structure. + * @param[in] item + * Item specification. + * @param[in] item_flags + * Parsed item flags. */ static void -flow_verbs_translate_item_vlan(const struct rte_flow_item *item, - uint64_t *item_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_item_vlan(struct mlx5_flow *dev_flow, + const struct rte_flow_item *item, + uint64_t item_flags) { const struct rte_flow_item_vlan *spec = item->spec; const struct rte_flow_item_vlan *mask = item->mask; unsigned int size = sizeof(struct ibv_flow_spec_eth); - const int tunnel = !!(*item_flags & MLX5_FLOW_LAYER_TUNNEL); + const int tunnel = !!(item_flags & MLX5_FLOW_LAYER_TUNNEL); struct ibv_flow_spec_eth eth = { - .type = IBV_FLOW_SPEC_ETH | (tunnel ? IBV_FLOW_SPEC_INNER : 0), + .type = IBV_FLOW_SPEC_ETH | VERBS_SPEC_INNER(item_flags), .size = size, }; const uint32_t l2m = tunnel ? MLX5_FLOW_LAYER_INNER_L2 : @@ -377,16 +522,13 @@ flow_verbs_translate_item_vlan(const struct rte_flow_item *item, eth.mask.ether_type = mask->inner_type; eth.val.ether_type &= eth.mask.ether_type; } - if (!(*item_flags & l2m)) { - dev_flow->verbs.attr->priority = MLX5_PRIORITY_MAP_L2; - flow_verbs_spec_add(dev_flow, ð, size); - } else { - flow_verbs_item_vlan_update(dev_flow->verbs.attr, ð); - size = 0; /* Only an update is done in eth specification. */ - } - *item_flags |= tunnel ? - (MLX5_FLOW_LAYER_INNER_L2 | MLX5_FLOW_LAYER_INNER_VLAN) : - (MLX5_FLOW_LAYER_OUTER_L2 | MLX5_FLOW_LAYER_OUTER_VLAN); + if (!(item_flags & l2m)) + flow_verbs_spec_add(&dev_flow->verbs, ð, size); + else + flow_verbs_item_vlan_update(&dev_flow->verbs.attr, ð); + if (!tunnel) + dev_flow->handle->vf_vlan.tag = + rte_be_to_cpu_16(spec->tci) & 0x0fff; } /** @@ -394,32 +536,28 @@ flow_verbs_translate_item_vlan(const struct rte_flow_item *item, * the input is valid and that there is space to insert the requested item * into the flow. * + * @param[in, out] dev_flow + * Pointer to dev_flow structure. * @param[in] item * Item specification. - * @param[in, out] item_flags - * Bit mask that marks all detected items. - * @param[in, out] dev_flow - * Pointer to sepacific flow structure. + * @param[in] item_flags + * Parsed item flags. */ static void -flow_verbs_translate_item_ipv4(const struct rte_flow_item *item, - uint64_t *item_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_item_ipv4(struct mlx5_flow *dev_flow, + const struct rte_flow_item *item, + uint64_t item_flags) { const struct rte_flow_item_ipv4 *spec = item->spec; const struct rte_flow_item_ipv4 *mask = item->mask; - const int tunnel = !!(*item_flags & MLX5_FLOW_LAYER_TUNNEL); unsigned int size = sizeof(struct ibv_flow_spec_ipv4_ext); struct ibv_flow_spec_ipv4_ext ipv4 = { - .type = IBV_FLOW_SPEC_IPV4_EXT | - (tunnel ? IBV_FLOW_SPEC_INNER : 0), + .type = IBV_FLOW_SPEC_IPV4_EXT | VERBS_SPEC_INNER(item_flags), .size = size, }; if (!mask) mask = &rte_flow_item_ipv4_mask; - *item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L3_IPV4 : - MLX5_FLOW_LAYER_OUTER_L3_IPV4; if (spec) { ipv4.val = (struct ibv_flow_ipv4_ext_filter){ .src_ip = spec->hdr.src_addr, @@ -439,12 +577,7 @@ flow_verbs_translate_item_ipv4(const struct rte_flow_item *item, ipv4.val.proto &= ipv4.mask.proto; ipv4.val.tos &= ipv4.mask.tos; } - dev_flow->verbs.hash_fields |= - mlx5_flow_hashfields_adjust(dev_flow, tunnel, - MLX5_IPV4_LAYER_TYPES, - MLX5_IPV4_IBV_RX_HASH); - dev_flow->verbs.attr->priority = MLX5_PRIORITY_MAP_L3; - flow_verbs_spec_add(dev_flow, &ipv4, size); + flow_verbs_spec_add(&dev_flow->verbs, &ipv4, size); } /** @@ -452,31 +585,28 @@ flow_verbs_translate_item_ipv4(const struct rte_flow_item *item, * the input is valid and that there is space to insert the requested item * into the flow. * + * @param[in, out] dev_flow + * Pointer to dev_flow structure. * @param[in] item * Item specification. - * @param[in, out] item_flags - * Bit mask that marks all detected items. - * @param[in, out] dev_flow - * Pointer to sepacific flow structure. + * @param[in] item_flags + * Parsed item flags. */ static void -flow_verbs_translate_item_ipv6(const struct rte_flow_item *item, - uint64_t *item_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_item_ipv6(struct mlx5_flow *dev_flow, + const struct rte_flow_item *item, + uint64_t item_flags) { const struct rte_flow_item_ipv6 *spec = item->spec; const struct rte_flow_item_ipv6 *mask = item->mask; - const int tunnel = !!(dev_flow->layers & MLX5_FLOW_LAYER_TUNNEL); unsigned int size = sizeof(struct ibv_flow_spec_ipv6); struct ibv_flow_spec_ipv6 ipv6 = { - .type = IBV_FLOW_SPEC_IPV6 | (tunnel ? IBV_FLOW_SPEC_INNER : 0), + .type = IBV_FLOW_SPEC_IPV6 | VERBS_SPEC_INNER(item_flags), .size = size, }; if (!mask) mask = &rte_flow_item_ipv6_mask; - *item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L3_IPV6 : - MLX5_FLOW_LAYER_OUTER_L3_IPV6; if (spec) { unsigned int i; uint32_t vtc_flow_val; @@ -493,19 +623,17 @@ flow_verbs_translate_item_ipv6(const struct rte_flow_item *item, vtc_flow_val = rte_be_to_cpu_32(spec->hdr.vtc_flow); vtc_flow_mask = rte_be_to_cpu_32(mask->hdr.vtc_flow); ipv6.val.flow_label = - rte_cpu_to_be_32((vtc_flow_val & IPV6_HDR_FL_MASK) >> - IPV6_HDR_FL_SHIFT); - ipv6.val.traffic_class = (vtc_flow_val & IPV6_HDR_TC_MASK) >> - IPV6_HDR_TC_SHIFT; + rte_cpu_to_be_32((vtc_flow_val & RTE_IPV6_HDR_FL_MASK) >> + RTE_IPV6_HDR_FL_SHIFT); + ipv6.val.traffic_class = (vtc_flow_val & RTE_IPV6_HDR_TC_MASK) >> + RTE_IPV6_HDR_TC_SHIFT; ipv6.val.next_hdr = spec->hdr.proto; - ipv6.val.hop_limit = spec->hdr.hop_limits; ipv6.mask.flow_label = - rte_cpu_to_be_32((vtc_flow_mask & IPV6_HDR_FL_MASK) >> - IPV6_HDR_FL_SHIFT); - ipv6.mask.traffic_class = (vtc_flow_mask & IPV6_HDR_TC_MASK) >> - IPV6_HDR_TC_SHIFT; + rte_cpu_to_be_32((vtc_flow_mask & RTE_IPV6_HDR_FL_MASK) >> + RTE_IPV6_HDR_FL_SHIFT); + ipv6.mask.traffic_class = (vtc_flow_mask & RTE_IPV6_HDR_TC_MASK) >> + RTE_IPV6_HDR_TC_SHIFT; ipv6.mask.next_hdr = mask->hdr.proto; - ipv6.mask.hop_limit = mask->hdr.hop_limits; /* Remove unwanted bits from values. */ for (i = 0; i < RTE_DIM(ipv6.val.src_ip); ++i) { ipv6.val.src_ip[i] &= ipv6.mask.src_ip[i]; @@ -514,14 +642,8 @@ flow_verbs_translate_item_ipv6(const struct rte_flow_item *item, ipv6.val.flow_label &= ipv6.mask.flow_label; ipv6.val.traffic_class &= ipv6.mask.traffic_class; ipv6.val.next_hdr &= ipv6.mask.next_hdr; - ipv6.val.hop_limit &= ipv6.mask.hop_limit; } - dev_flow->verbs.hash_fields |= - mlx5_flow_hashfields_adjust(dev_flow, tunnel, - MLX5_IPV6_LAYER_TYPES, - MLX5_IPV6_IBV_RX_HASH); - dev_flow->verbs.attr->priority = MLX5_PRIORITY_MAP_L3; - flow_verbs_spec_add(dev_flow, &ipv6, size); + flow_verbs_spec_add(&dev_flow->verbs, &ipv6, size); } /** @@ -529,46 +651,38 @@ flow_verbs_translate_item_ipv6(const struct rte_flow_item *item, * the input is valid and that there is space to insert the requested item * into the flow. * + * @param[in, out] dev_flow + * Pointer to dev_flow structure. * @param[in] item * Item specification. - * @param[in, out] item_flags - * Bit mask that marks all detected items. - * @param[in, out] dev_flow - * Pointer to sepacific flow structure. + * @param[in] item_flags + * Parsed item flags. */ static void -flow_verbs_translate_item_udp(const struct rte_flow_item *item, - uint64_t *item_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_item_tcp(struct mlx5_flow *dev_flow, + const struct rte_flow_item *item, + uint64_t item_flags __rte_unused) { - const struct rte_flow_item_udp *spec = item->spec; - const struct rte_flow_item_udp *mask = item->mask; - const int tunnel = !!(*item_flags & MLX5_FLOW_LAYER_TUNNEL); + const struct rte_flow_item_tcp *spec = item->spec; + const struct rte_flow_item_tcp *mask = item->mask; unsigned int size = sizeof(struct ibv_flow_spec_tcp_udp); - struct ibv_flow_spec_tcp_udp udp = { - .type = IBV_FLOW_SPEC_UDP | (tunnel ? IBV_FLOW_SPEC_INNER : 0), + struct ibv_flow_spec_tcp_udp tcp = { + .type = IBV_FLOW_SPEC_TCP | VERBS_SPEC_INNER(item_flags), .size = size, }; if (!mask) - mask = &rte_flow_item_udp_mask; - *item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L4_UDP : - MLX5_FLOW_LAYER_OUTER_L4_UDP; + mask = &rte_flow_item_tcp_mask; if (spec) { - udp.val.dst_port = spec->hdr.dst_port; - udp.val.src_port = spec->hdr.src_port; - udp.mask.dst_port = mask->hdr.dst_port; - udp.mask.src_port = mask->hdr.src_port; + tcp.val.dst_port = spec->hdr.dst_port; + tcp.val.src_port = spec->hdr.src_port; + tcp.mask.dst_port = mask->hdr.dst_port; + tcp.mask.src_port = mask->hdr.src_port; /* Remove unwanted bits from values. */ - udp.val.src_port &= udp.mask.src_port; - udp.val.dst_port &= udp.mask.dst_port; + tcp.val.src_port &= tcp.mask.src_port; + tcp.val.dst_port &= tcp.mask.dst_port; } - dev_flow->verbs.hash_fields |= - mlx5_flow_hashfields_adjust(dev_flow, tunnel, ETH_RSS_UDP, - (IBV_RX_HASH_SRC_PORT_UDP | - IBV_RX_HASH_DST_PORT_UDP)); - dev_flow->verbs.attr->priority = MLX5_PRIORITY_MAP_L4; - flow_verbs_spec_add(dev_flow, &udp, size); + flow_verbs_spec_add(&dev_flow->verbs, &tcp, size); } /** @@ -576,46 +690,60 @@ flow_verbs_translate_item_udp(const struct rte_flow_item *item, * the input is valid and that there is space to insert the requested item * into the flow. * + * @param[in, out] dev_flow + * Pointer to dev_flow structure. * @param[in] item * Item specification. - * @param[in, out] item_flags - * Bit mask that marks all detected items. - * @param[in, out] dev_flow - * Pointer to sepacific flow structure. + * @param[in] item_flags + * Parsed item flags. */ static void -flow_verbs_translate_item_tcp(const struct rte_flow_item *item, - uint64_t *item_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_item_udp(struct mlx5_flow *dev_flow, + const struct rte_flow_item *item, + uint64_t item_flags __rte_unused) { - const struct rte_flow_item_tcp *spec = item->spec; - const struct rte_flow_item_tcp *mask = item->mask; - const int tunnel = !!(dev_flow->layers & MLX5_FLOW_LAYER_TUNNEL); + const struct rte_flow_item_udp *spec = item->spec; + const struct rte_flow_item_udp *mask = item->mask; unsigned int size = sizeof(struct ibv_flow_spec_tcp_udp); - struct ibv_flow_spec_tcp_udp tcp = { - .type = IBV_FLOW_SPEC_TCP | (tunnel ? IBV_FLOW_SPEC_INNER : 0), + struct ibv_flow_spec_tcp_udp udp = { + .type = IBV_FLOW_SPEC_UDP | VERBS_SPEC_INNER(item_flags), .size = size, }; if (!mask) - mask = &rte_flow_item_tcp_mask; - *item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L4_TCP : - MLX5_FLOW_LAYER_OUTER_L4_TCP; + mask = &rte_flow_item_udp_mask; if (spec) { - tcp.val.dst_port = spec->hdr.dst_port; - tcp.val.src_port = spec->hdr.src_port; - tcp.mask.dst_port = mask->hdr.dst_port; - tcp.mask.src_port = mask->hdr.src_port; + udp.val.dst_port = spec->hdr.dst_port; + udp.val.src_port = spec->hdr.src_port; + udp.mask.dst_port = mask->hdr.dst_port; + udp.mask.src_port = mask->hdr.src_port; /* Remove unwanted bits from values. */ - tcp.val.src_port &= tcp.mask.src_port; - tcp.val.dst_port &= tcp.mask.dst_port; + udp.val.src_port &= udp.mask.src_port; + udp.val.dst_port &= udp.mask.dst_port; } - dev_flow->verbs.hash_fields |= - mlx5_flow_hashfields_adjust(dev_flow, tunnel, ETH_RSS_TCP, - (IBV_RX_HASH_SRC_PORT_TCP | - IBV_RX_HASH_DST_PORT_TCP)); - dev_flow->verbs.attr->priority = MLX5_PRIORITY_MAP_L4; - flow_verbs_spec_add(dev_flow, &tcp, size); + item++; + while (item->type == RTE_FLOW_ITEM_TYPE_VOID) + item++; + if (!(udp.val.dst_port & udp.mask.dst_port)) { + switch ((item)->type) { + case RTE_FLOW_ITEM_TYPE_VXLAN: + udp.val.dst_port = htons(MLX5_UDP_PORT_VXLAN); + udp.mask.dst_port = 0xffff; + break; + case RTE_FLOW_ITEM_TYPE_VXLAN_GPE: + udp.val.dst_port = htons(MLX5_UDP_PORT_VXLAN_GPE); + udp.mask.dst_port = 0xffff; + break; + case RTE_FLOW_ITEM_TYPE_MPLS: + udp.val.dst_port = htons(MLX5_UDP_PORT_MPLS); + udp.mask.dst_port = 0xffff; + break; + default: + break; + } + } + + flow_verbs_spec_add(&dev_flow->verbs, &udp, size); } /** @@ -623,17 +751,17 @@ flow_verbs_translate_item_tcp(const struct rte_flow_item *item, * the input is valid and that there is space to insert the requested item * into the flow. * + * @param[in, out] dev_flow + * Pointer to dev_flow structure. * @param[in] item * Item specification. - * @param[in, out] item_flags - * Bit mask that marks all detected items. - * @param[in, out] dev_flow - * Pointer to sepacific flow structure. + * @param[in] item_flags + * Parsed item flags. */ static void -flow_verbs_translate_item_vxlan(const struct rte_flow_item *item, - uint64_t *item_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_item_vxlan(struct mlx5_flow *dev_flow, + const struct rte_flow_item *item, + uint64_t item_flags __rte_unused) { const struct rte_flow_item_vxlan *spec = item->spec; const struct rte_flow_item_vxlan *mask = item->mask; @@ -657,9 +785,7 @@ flow_verbs_translate_item_vxlan(const struct rte_flow_item *item, /* Remove unwanted bits from values. */ vxlan.val.tunnel_id &= vxlan.mask.tunnel_id; } - flow_verbs_spec_add(dev_flow, &vxlan, size); - dev_flow->verbs.attr->priority = MLX5_PRIORITY_MAP_L2; - *item_flags |= MLX5_FLOW_LAYER_VXLAN; + flow_verbs_spec_add(&dev_flow->verbs, &vxlan, size); } /** @@ -667,17 +793,17 @@ flow_verbs_translate_item_vxlan(const struct rte_flow_item *item, * the input is valid and that there is space to insert the requested item * into the flow. * + * @param[in, out] dev_flow + * Pointer to dev_flow structure. * @param[in] item * Item specification. - * @param[in, out] item_flags - * Bit mask that marks all detected items. - * @param[in, out] dev_flow - * Pointer to sepacific flow structure. + * @param[in] item_flags + * Parsed item flags. */ static void -flow_verbs_translate_item_vxlan_gpe(const struct rte_flow_item *item, - uint64_t *item_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_item_vxlan_gpe(struct mlx5_flow *dev_flow, + const struct rte_flow_item *item, + uint64_t item_flags __rte_unused) { const struct rte_flow_item_vxlan_gpe *spec = item->spec; const struct rte_flow_item_vxlan_gpe *mask = item->mask; @@ -701,9 +827,7 @@ flow_verbs_translate_item_vxlan_gpe(const struct rte_flow_item *item, /* Remove unwanted bits from values. */ vxlan_gpe.val.tunnel_id &= vxlan_gpe.mask.tunnel_id; } - flow_verbs_spec_add(dev_flow, &vxlan_gpe, size); - dev_flow->verbs.attr->priority = MLX5_PRIORITY_MAP_L2; - *item_flags |= MLX5_FLOW_LAYER_VXLAN_GPE; + flow_verbs_spec_add(&dev_flow->verbs, &vxlan_gpe, size); } /** @@ -763,19 +887,19 @@ flow_verbs_item_gre_ip_protocol_update(struct ibv_flow_attr *attr, * the input is valid and that there is space to insert the requested item * into the flow. * + * @param[in, out] dev_flow + * Pointer to dev_flow structure. * @param[in] item * Item specification. - * @param[in, out] item_flags - * Bit mask that marks all detected items. - * @param[in, out] dev_flow - * Pointer to sepacific flow structure. + * @param[in] item_flags + * Parsed item flags. */ static void -flow_verbs_translate_item_gre(const struct rte_flow_item *item __rte_unused, - uint64_t *item_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_item_gre(struct mlx5_flow *dev_flow, + const struct rte_flow_item *item __rte_unused, + uint64_t item_flags) { - struct mlx5_flow_verbs *verbs = &dev_flow->verbs; + struct mlx5_flow_verbs_workspace *verbs = &dev_flow->verbs; #ifndef HAVE_IBV_DEVICE_MPLS_SUPPORT unsigned int size = sizeof(struct ibv_flow_spec_tunnel); struct ibv_flow_spec_tunnel tunnel = { @@ -804,17 +928,15 @@ flow_verbs_translate_item_gre(const struct rte_flow_item *item __rte_unused, tunnel.val.key &= tunnel.mask.key; } #endif - if (*item_flags & MLX5_FLOW_LAYER_OUTER_L3_IPV4) - flow_verbs_item_gre_ip_protocol_update(verbs->attr, + if (item_flags & MLX5_FLOW_LAYER_OUTER_L3_IPV4) + flow_verbs_item_gre_ip_protocol_update(&verbs->attr, IBV_FLOW_SPEC_IPV4_EXT, IPPROTO_GRE); else - flow_verbs_item_gre_ip_protocol_update(verbs->attr, + flow_verbs_item_gre_ip_protocol_update(&verbs->attr, IBV_FLOW_SPEC_IPV6, IPPROTO_GRE); - flow_verbs_spec_add(dev_flow, &tunnel, size); - verbs->attr->priority = MLX5_PRIORITY_MAP_L2; - *item_flags |= MLX5_FLOW_LAYER_GRE; + flow_verbs_spec_add(verbs, &tunnel, size); } /** @@ -822,17 +944,17 @@ flow_verbs_translate_item_gre(const struct rte_flow_item *item __rte_unused, * the input is valid and that there is space to insert the requested action * into the flow. This function also return the action that was added. * + * @param[in, out] dev_flow + * Pointer to dev_flow structure. * @param[in] item * Item specification. - * @param[in, out] item_flags - * Bit mask that marks all detected items. - * @param[in, out] dev_flow - * Pointer to sepacific flow structure. + * @param[in] item_flags + * Parsed item flags. */ static void -flow_verbs_translate_item_mpls(const struct rte_flow_item *item __rte_unused, - uint64_t *action_flags __rte_unused, - struct mlx5_flow *dev_flow __rte_unused) +flow_verbs_translate_item_mpls(struct mlx5_flow *dev_flow __rte_unused, + const struct rte_flow_item *item __rte_unused, + uint64_t item_flags __rte_unused) { #ifdef HAVE_IBV_DEVICE_MPLS_SUPPORT const struct rte_flow_item_mpls *spec = item->spec; @@ -851,25 +973,24 @@ flow_verbs_translate_item_mpls(const struct rte_flow_item *item __rte_unused, /* Remove unwanted bits from values. */ mpls.val.label &= mpls.mask.label; } - flow_verbs_spec_add(dev_flow, &mpls, size); - dev_flow->verbs.attr->priority = MLX5_PRIORITY_MAP_L2; - *action_flags |= MLX5_FLOW_LAYER_MPLS; + flow_verbs_spec_add(&dev_flow->verbs, &mpls, size); #endif } /** * Convert the @p action into a Verbs specification. This function assumes that * the input is valid and that there is space to insert the requested action - * into the flow. This function also return the action that was added. + * into the flow. * - * @param[in, out] action_flags - * Pointer to the detected actions. * @param[in] dev_flow * Pointer to mlx5_flow. + * @param[in] action + * Action configuration. */ static void -flow_verbs_translate_action_drop(uint64_t *action_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_action_drop + (struct mlx5_flow *dev_flow, + const struct rte_flow_action *action __rte_unused) { unsigned int size = sizeof(struct ibv_flow_spec_action_drop); struct ibv_flow_spec_action_drop drop = { @@ -877,83 +998,71 @@ flow_verbs_translate_action_drop(uint64_t *action_flags, .size = size, }; - flow_verbs_spec_add(dev_flow, &drop, size); - *action_flags |= MLX5_FLOW_ACTION_DROP; + flow_verbs_spec_add(&dev_flow->verbs, &drop, size); } /** * Convert the @p action into a Verbs specification. This function assumes that * the input is valid and that there is space to insert the requested action - * into the flow. This function also return the action that was added. + * into the flow. * + * @param[in] rss_desc + * Pointer to mlx5_flow_rss_desc. * @param[in] action * Action configuration. - * @param[in, out] action_flags - * Pointer to the detected actions. - * @param[in] dev_flow - * Pointer to mlx5_flow. */ static void -flow_verbs_translate_action_queue(const struct rte_flow_action *action, - uint64_t *action_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_action_queue(struct mlx5_flow_rss_desc *rss_desc, + const struct rte_flow_action *action) { const struct rte_flow_action_queue *queue = action->conf; - struct rte_flow *flow = dev_flow->flow; - if (flow->queue) - (*flow->queue)[0] = queue->index; - flow->rss.queue_num = 1; - *action_flags |= MLX5_FLOW_ACTION_QUEUE; + rss_desc->queue[0] = queue->index; + rss_desc->queue_num = 1; } /** * Convert the @p action into a Verbs specification. This function assumes that * the input is valid and that there is space to insert the requested action - * into the flow. This function also return the action that was added. + * into the flow. * + * @param[in] rss_desc + * Pointer to mlx5_flow_rss_desc. * @param[in] action * Action configuration. - * @param[in, out] action_flags - * Pointer to the detected actions. - * @param[in] dev_flow - * Pointer to mlx5_flow. */ static void -flow_verbs_translate_action_rss(const struct rte_flow_action *action, - uint64_t *action_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_action_rss(struct mlx5_flow_rss_desc *rss_desc, + const struct rte_flow_action *action) { const struct rte_flow_action_rss *rss = action->conf; - struct rte_flow *flow = dev_flow->flow; - - if (flow->queue) - memcpy((*flow->queue), rss->queue, - rss->queue_num * sizeof(uint16_t)); - flow->rss.queue_num = rss->queue_num; - memcpy(flow->key, rss->key, MLX5_RSS_HASH_KEY_LEN); - flow->rss.types = rss->types; - flow->rss.level = rss->level; - *action_flags |= MLX5_FLOW_ACTION_RSS; + const uint8_t *rss_key; + + memcpy(rss_desc->queue, rss->queue, rss->queue_num * sizeof(uint16_t)); + rss_desc->queue_num = rss->queue_num; + /* NULL RSS key indicates default RSS key. */ + rss_key = !rss->key ? rss_hash_default_key : rss->key; + memcpy(rss_desc->key, rss_key, MLX5_RSS_HASH_KEY_LEN); + /* + * rss->level and rss.types should be set in advance when expanding + * items for RSS. + */ } /** * Convert the @p action into a Verbs specification. This function assumes that * the input is valid and that there is space to insert the requested action - * into the flow. This function also return the action that was added. + * into the flow. * - * @param[in] action - * Action configuration. - * @param[in, out] action_flags - * Pointer to the detected actions. * @param[in] dev_flow * Pointer to mlx5_flow. + * @param[in] action + * Action configuration. */ static void flow_verbs_translate_action_flag - (const struct rte_flow_action *action __rte_unused, - uint64_t *action_flags, - struct mlx5_flow *dev_flow) + (struct mlx5_flow *dev_flow, + const struct rte_flow_action *action __rte_unused) { unsigned int size = sizeof(struct ibv_flow_spec_action_tag); struct ibv_flow_spec_action_tag tag = { @@ -961,87 +1070,44 @@ flow_verbs_translate_action_flag .size = size, .tag_id = mlx5_flow_mark_set(MLX5_FLOW_MARK_DEFAULT), }; - *action_flags |= MLX5_FLOW_ACTION_MARK; - flow_verbs_spec_add(dev_flow, &tag, size); -} - -/** - * Update verbs specification to modify the flag to mark. - * - * @param[in, out] verbs - * Pointer to the mlx5_flow_verbs structure. - * @param[in] mark_id - * Mark identifier to replace the flag. - */ -static void -flow_verbs_mark_update(struct mlx5_flow_verbs *verbs, uint32_t mark_id) -{ - struct ibv_spec_header *hdr; - int i; - - if (!verbs) - return; - /* Update Verbs specification. */ - hdr = (struct ibv_spec_header *)verbs->specs; - if (!hdr) - return; - for (i = 0; i != verbs->attr->num_of_specs; ++i) { - if (hdr->type == IBV_FLOW_SPEC_ACTION_TAG) { - struct ibv_flow_spec_action_tag *t = - (struct ibv_flow_spec_action_tag *)hdr; - t->tag_id = mlx5_flow_mark_set(mark_id); - } - hdr = (struct ibv_spec_header *)((uintptr_t)hdr + hdr->size); - } + flow_verbs_spec_add(&dev_flow->verbs, &tag, size); } /** * Convert the @p action into a Verbs specification. This function assumes that * the input is valid and that there is space to insert the requested action - * into the flow. This function also return the action that was added. + * into the flow. * - * @param[in] action - * Action configuration. - * @param[in, out] action_flags - * Pointer to the detected actions. * @param[in] dev_flow * Pointer to mlx5_flow. + * @param[in] action + * Action configuration. */ static void -flow_verbs_translate_action_mark(const struct rte_flow_action *action, - uint64_t *action_flags, - struct mlx5_flow *dev_flow) +flow_verbs_translate_action_mark(struct mlx5_flow *dev_flow, + const struct rte_flow_action *action) { const struct rte_flow_action_mark *mark = action->conf; unsigned int size = sizeof(struct ibv_flow_spec_action_tag); struct ibv_flow_spec_action_tag tag = { .type = IBV_FLOW_SPEC_ACTION_TAG, .size = size, + .tag_id = mlx5_flow_mark_set(mark->id), }; - struct mlx5_flow_verbs *verbs = &dev_flow->verbs; - - if (*action_flags & MLX5_FLOW_ACTION_FLAG) { - flow_verbs_mark_update(verbs, mark->id); - size = 0; - } else { - tag.tag_id = mlx5_flow_mark_set(mark->id); - flow_verbs_spec_add(dev_flow, &tag, size); - } - *action_flags |= MLX5_FLOW_ACTION_MARK; + + flow_verbs_spec_add(&dev_flow->verbs, &tag, size); } /** * Convert the @p action into a Verbs specification. This function assumes that * the input is valid and that there is space to insert the requested action - * into the flow. This function also return the action that was added. + * into the flow. * * @param[in] dev * Pointer to the Ethernet device structure. * @param[in] action * Action configuration. - * @param[in, out] action_flags - * Pointer to the detected actions. * @param[in] dev_flow * Pointer to mlx5_flow. * @param[out] error @@ -1051,16 +1117,17 @@ flow_verbs_translate_action_mark(const struct rte_flow_action *action, * 0 On success else a negative errno value is returned and rte_errno is set. */ static int -flow_verbs_translate_action_count(struct rte_eth_dev *dev, +flow_verbs_translate_action_count(struct mlx5_flow *dev_flow, const struct rte_flow_action *action, - uint64_t *action_flags, - struct mlx5_flow *dev_flow, + struct rte_eth_dev *dev, struct rte_flow_error *error) { const struct rte_flow_action_count *count = action->conf; struct rte_flow *flow = dev_flow->flow; #if defined(HAVE_IBV_DEVICE_COUNTERS_SET_V42) || \ defined(HAVE_IBV_DEVICE_COUNTERS_SET_V45) + struct mlx5_flow_counter_pool *pool; + struct mlx5_flow_counter *cnt = NULL; unsigned int size = sizeof(struct ibv_flow_spec_counter_action); struct ibv_flow_spec_counter_action counter = { .type = IBV_FLOW_SPEC_ACTION_COUNT, @@ -1069,8 +1136,7 @@ flow_verbs_translate_action_count(struct rte_eth_dev *dev, #endif if (!flow->counter) { - flow->counter = flow_verbs_counter_new(dev, count->shared, - count->id); + flow->counter = flow_verbs_counter_new(dev, count->id); if (!flow->counter) return rte_flow_error_set(error, rte_errno, RTE_FLOW_ERROR_TYPE_ACTION, @@ -1078,13 +1144,15 @@ flow_verbs_translate_action_count(struct rte_eth_dev *dev, "cannot get counter" " context."); } - *action_flags |= MLX5_FLOW_ACTION_COUNT; #if defined(HAVE_IBV_DEVICE_COUNTERS_SET_V42) - counter.counter_set_handle = flow->counter->cs->handle; - flow_verbs_spec_add(dev_flow, &counter, size); + cnt = flow_verbs_counter_get_by_idx(dev, flow->counter, &pool); + counter.counter_set_handle = + ((struct ibv_counter_set *)cnt->dcs_when_active)->handle; + flow_verbs_spec_add(&dev_flow->verbs, &counter, size); #elif defined(HAVE_IBV_DEVICE_COUNTERS_SET_V45) - counter.counters = flow->counter->cs; - flow_verbs_spec_add(dev_flow, &counter, size); + cnt = flow_verbs_counter_get_by_idx(dev, flow->counter, &pool); + counter.counters = (struct ibv_counters *)cnt->dcs_when_active; + flow_verbs_spec_add(&dev_flow->verbs, &counter, size); #endif return 0; } @@ -1100,6 +1168,10 @@ flow_verbs_translate_action_count(struct rte_eth_dev *dev, * Pointer to the list of items. * @param[in] actions * Pointer to the list of actions. + * @param[in] external + * This flow rule is created by request external to PMD. + * @param[in] hairpin + * Number of hairpin TX actions, 0 means classic flow. * @param[out] error * Pointer to the error structure. * @@ -1111,13 +1183,18 @@ flow_verbs_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, const struct rte_flow_item items[], const struct rte_flow_action actions[], + bool external __rte_unused, + int hairpin __rte_unused, struct rte_flow_error *error) { int ret; - uint32_t action_flags = 0; - uint32_t item_flags = 0; - int tunnel = 0; + uint64_t action_flags = 0; + uint64_t item_flags = 0; + uint64_t last_item = 0; uint8_t next_protocol = 0xff; + uint16_t ether_type = 0; + bool is_empty_vlan = false; + uint16_t udp_dport = 0; if (items == NULL) return -1; @@ -1125,64 +1202,119 @@ flow_verbs_validate(struct rte_eth_dev *dev, if (ret < 0) return ret; for (; items->type != RTE_FLOW_ITEM_TYPE_END; items++) { + int tunnel = !!(item_flags & MLX5_FLOW_LAYER_TUNNEL); int ret = 0; - tunnel = !!(item_flags & MLX5_FLOW_LAYER_TUNNEL); switch (items->type) { case RTE_FLOW_ITEM_TYPE_VOID: break; case RTE_FLOW_ITEM_TYPE_ETH: ret = mlx5_flow_validate_item_eth(items, item_flags, - error); + false, error); if (ret < 0) return ret; - item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L2 : - MLX5_FLOW_LAYER_OUTER_L2; + last_item = tunnel ? MLX5_FLOW_LAYER_INNER_L2 : + MLX5_FLOW_LAYER_OUTER_L2; + if (items->mask != NULL && items->spec != NULL) { + ether_type = + ((const struct rte_flow_item_eth *) + items->spec)->type; + ether_type &= + ((const struct rte_flow_item_eth *) + items->mask)->type; + if (ether_type == RTE_BE16(RTE_ETHER_TYPE_VLAN)) + is_empty_vlan = true; + ether_type = rte_be_to_cpu_16(ether_type); + } else { + ether_type = 0; + } break; case RTE_FLOW_ITEM_TYPE_VLAN: ret = mlx5_flow_validate_item_vlan(items, item_flags, - error); + dev, error); if (ret < 0) return ret; - item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_VLAN : - MLX5_FLOW_LAYER_OUTER_VLAN; + last_item = tunnel ? (MLX5_FLOW_LAYER_INNER_L2 | + MLX5_FLOW_LAYER_INNER_VLAN) : + (MLX5_FLOW_LAYER_OUTER_L2 | + MLX5_FLOW_LAYER_OUTER_VLAN); + if (items->mask != NULL && items->spec != NULL) { + ether_type = + ((const struct rte_flow_item_vlan *) + items->spec)->inner_type; + ether_type &= + ((const struct rte_flow_item_vlan *) + items->mask)->inner_type; + ether_type = rte_be_to_cpu_16(ether_type); + } else { + ether_type = 0; + } + is_empty_vlan = false; break; case RTE_FLOW_ITEM_TYPE_IPV4: - ret = mlx5_flow_validate_item_ipv4(items, item_flags, - error); + ret = mlx5_flow_validate_item_ipv4 + (items, item_flags, + last_item, ether_type, NULL, + MLX5_ITEM_RANGE_NOT_ACCEPTED, + error); if (ret < 0) return ret; - item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L3_IPV4 : - MLX5_FLOW_LAYER_OUTER_L3_IPV4; + last_item = tunnel ? MLX5_FLOW_LAYER_INNER_L3_IPV4 : + MLX5_FLOW_LAYER_OUTER_L3_IPV4; if (items->mask != NULL && ((const struct rte_flow_item_ipv4 *) - items->mask)->hdr.next_proto_id) + items->mask)->hdr.next_proto_id) { next_protocol = ((const struct rte_flow_item_ipv4 *) (items->spec))->hdr.next_proto_id; + next_protocol &= + ((const struct rte_flow_item_ipv4 *) + (items->mask))->hdr.next_proto_id; + } else { + /* Reset for inner layer. */ + next_protocol = 0xff; + } break; case RTE_FLOW_ITEM_TYPE_IPV6: ret = mlx5_flow_validate_item_ipv6(items, item_flags, + last_item, + ether_type, NULL, error); if (ret < 0) return ret; - item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L3_IPV6 : - MLX5_FLOW_LAYER_OUTER_L3_IPV6; + last_item = tunnel ? MLX5_FLOW_LAYER_INNER_L3_IPV6 : + MLX5_FLOW_LAYER_OUTER_L3_IPV6; if (items->mask != NULL && ((const struct rte_flow_item_ipv6 *) - items->mask)->hdr.proto) + items->mask)->hdr.proto) { next_protocol = ((const struct rte_flow_item_ipv6 *) items->spec)->hdr.proto; + next_protocol &= + ((const struct rte_flow_item_ipv6 *) + items->mask)->hdr.proto; + } else { + /* Reset for inner layer. */ + next_protocol = 0xff; + } break; case RTE_FLOW_ITEM_TYPE_UDP: ret = mlx5_flow_validate_item_udp(items, item_flags, next_protocol, error); + const struct rte_flow_item_udp *spec = items->spec; + const struct rte_flow_item_udp *mask = items->mask; + if (!mask) + mask = &rte_flow_item_udp_mask; + if (spec != NULL) + udp_dport = rte_be_to_cpu_16 + (spec->hdr.dst_port & + mask->hdr.dst_port); + if (ret < 0) return ret; - item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L4_UDP : - MLX5_FLOW_LAYER_OUTER_L4_UDP; + last_item = tunnel ? MLX5_FLOW_LAYER_INNER_L4_UDP : + MLX5_FLOW_LAYER_OUTER_L4_UDP; break; case RTE_FLOW_ITEM_TYPE_TCP: ret = mlx5_flow_validate_item_tcp @@ -1192,15 +1324,16 @@ flow_verbs_validate(struct rte_eth_dev *dev, error); if (ret < 0) return ret; - item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L4_TCP : - MLX5_FLOW_LAYER_OUTER_L4_TCP; + last_item = tunnel ? MLX5_FLOW_LAYER_INNER_L4_TCP : + MLX5_FLOW_LAYER_OUTER_L4_TCP; break; case RTE_FLOW_ITEM_TYPE_VXLAN: - ret = mlx5_flow_validate_item_vxlan(items, item_flags, - error); + ret = mlx5_flow_validate_item_vxlan(dev, udp_dport, + items, item_flags, + attr, error); if (ret < 0) return ret; - item_flags |= MLX5_FLOW_LAYER_VXLAN; + last_item = MLX5_FLOW_LAYER_VXLAN; break; case RTE_FLOW_ITEM_TYPE_VXLAN_GPE: ret = mlx5_flow_validate_item_vxlan_gpe(items, @@ -1208,36 +1341,40 @@ flow_verbs_validate(struct rte_eth_dev *dev, dev, error); if (ret < 0) return ret; - item_flags |= MLX5_FLOW_LAYER_VXLAN_GPE; + last_item = MLX5_FLOW_LAYER_VXLAN_GPE; break; case RTE_FLOW_ITEM_TYPE_GRE: ret = mlx5_flow_validate_item_gre(items, item_flags, next_protocol, error); if (ret < 0) return ret; - item_flags |= MLX5_FLOW_LAYER_GRE; + last_item = MLX5_FLOW_LAYER_GRE; break; case RTE_FLOW_ITEM_TYPE_MPLS: - ret = mlx5_flow_validate_item_mpls(items, item_flags, - next_protocol, - error); + ret = mlx5_flow_validate_item_mpls(dev, items, + item_flags, + last_item, error); if (ret < 0) return ret; - if (next_protocol != 0xff && - next_protocol != IPPROTO_MPLS) - return rte_flow_error_set - (error, EINVAL, - RTE_FLOW_ERROR_TYPE_ITEM, items, - "protocol filtering not compatible" - " with MPLS layer"); - item_flags |= MLX5_FLOW_LAYER_MPLS; + last_item = MLX5_FLOW_LAYER_MPLS; break; + case RTE_FLOW_ITEM_TYPE_ICMP: + case RTE_FLOW_ITEM_TYPE_ICMP6: + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ITEM, + NULL, "ICMP/ICMP6 " + "item not supported"); default: return rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM, NULL, "item not supported"); } + item_flags |= last_item; } + if (is_empty_vlan) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ITEM, NULL, + "VLAN matching without vid specification is not supported"); for (; actions->type != RTE_FLOW_ACTION_TYPE_END; actions++) { switch (actions->type) { case RTE_FLOW_ACTION_TYPE_VOID: @@ -1279,7 +1416,7 @@ flow_verbs_validate(struct rte_eth_dev *dev, case RTE_FLOW_ACTION_TYPE_RSS: ret = mlx5_flow_validate_action_rss(actions, action_flags, dev, - attr, + attr, item_flags, error); if (ret < 0) return ret; @@ -1298,6 +1435,18 @@ flow_verbs_validate(struct rte_eth_dev *dev, "action not supported"); } } + /* + * Validate the drop action mutual exclusion with other actions. + * Drop action is mutually-exclusive with any other action, except for + * Count action. + */ + if ((action_flags & MLX5_FLOW_ACTION_DROP) && + (action_flags & ~(MLX5_FLOW_ACTION_DROP | MLX5_FLOW_ACTION_COUNT))) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, NULL, + "Drop action is mutually-exclusive " + "with any other action, except for " + "Count action"); if (!(action_flags & MLX5_FLOW_FATE_ACTIONS)) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, actions, @@ -1307,23 +1456,18 @@ flow_verbs_validate(struct rte_eth_dev *dev, /** * Calculate the required bytes that are needed for the action part of the verbs - * flow, in addtion returns bit-fields with all the detected action, in order to - * avoid another interation over the actions. + * flow. * * @param[in] actions * Pointer to the list of actions. - * @param[out] action_flags - * Pointer to the detected actions. * * @return * The size of the memory needed for all actions. */ static int -flow_verbs_get_actions_and_size(const struct rte_flow_action actions[], - uint64_t *action_flags) +flow_verbs_get_actions_size(const struct rte_flow_action actions[]) { int size = 0; - uint64_t detected_actions = 0; for (; actions->type != RTE_FLOW_ACTION_TYPE_END; actions++) { switch (actions->type) { @@ -1331,125 +1475,89 @@ flow_verbs_get_actions_and_size(const struct rte_flow_action actions[], break; case RTE_FLOW_ACTION_TYPE_FLAG: size += sizeof(struct ibv_flow_spec_action_tag); - detected_actions |= MLX5_FLOW_ACTION_FLAG; break; case RTE_FLOW_ACTION_TYPE_MARK: size += sizeof(struct ibv_flow_spec_action_tag); - detected_actions |= MLX5_FLOW_ACTION_MARK; break; case RTE_FLOW_ACTION_TYPE_DROP: size += sizeof(struct ibv_flow_spec_action_drop); - detected_actions |= MLX5_FLOW_ACTION_DROP; break; case RTE_FLOW_ACTION_TYPE_QUEUE: - detected_actions |= MLX5_FLOW_ACTION_QUEUE; break; case RTE_FLOW_ACTION_TYPE_RSS: - detected_actions |= MLX5_FLOW_ACTION_RSS; break; case RTE_FLOW_ACTION_TYPE_COUNT: #if defined(HAVE_IBV_DEVICE_COUNTERS_SET_V42) || \ defined(HAVE_IBV_DEVICE_COUNTERS_SET_V45) size += sizeof(struct ibv_flow_spec_counter_action); #endif - detected_actions |= MLX5_FLOW_ACTION_COUNT; break; default: break; } } - *action_flags = detected_actions; return size; } /** * Calculate the required bytes that are needed for the item part of the verbs - * flow, in addtion returns bit-fields with all the detected action, in order to - * avoid another interation over the actions. + * flow. * - * @param[in] actions + * @param[in] items * Pointer to the list of items. - * @param[in, out] item_flags - * Pointer to the detected items. * * @return * The size of the memory needed for all items. */ static int -flow_verbs_get_items_and_size(const struct rte_flow_item items[], - uint64_t *item_flags) +flow_verbs_get_items_size(const struct rte_flow_item items[]) { int size = 0; - uint64_t detected_items = 0; for (; items->type != RTE_FLOW_ITEM_TYPE_END; items++) { - int tunnel = !!(detected_items & MLX5_FLOW_LAYER_TUNNEL); - switch (items->type) { case RTE_FLOW_ITEM_TYPE_VOID: break; case RTE_FLOW_ITEM_TYPE_ETH: size += sizeof(struct ibv_flow_spec_eth); - detected_items |= tunnel ? MLX5_FLOW_LAYER_INNER_L2 : - MLX5_FLOW_LAYER_OUTER_L2; break; case RTE_FLOW_ITEM_TYPE_VLAN: size += sizeof(struct ibv_flow_spec_eth); - detected_items |= tunnel ? MLX5_FLOW_LAYER_INNER_VLAN : - MLX5_FLOW_LAYER_OUTER_VLAN; break; case RTE_FLOW_ITEM_TYPE_IPV4: size += sizeof(struct ibv_flow_spec_ipv4_ext); - detected_items |= tunnel ? - MLX5_FLOW_LAYER_INNER_L3_IPV4 : - MLX5_FLOW_LAYER_OUTER_L3_IPV4; break; case RTE_FLOW_ITEM_TYPE_IPV6: size += sizeof(struct ibv_flow_spec_ipv6); - detected_items |= tunnel ? - MLX5_FLOW_LAYER_INNER_L3_IPV6 : - MLX5_FLOW_LAYER_OUTER_L3_IPV6; break; case RTE_FLOW_ITEM_TYPE_UDP: size += sizeof(struct ibv_flow_spec_tcp_udp); - detected_items |= tunnel ? - MLX5_FLOW_LAYER_INNER_L4_UDP : - MLX5_FLOW_LAYER_OUTER_L4_UDP; break; case RTE_FLOW_ITEM_TYPE_TCP: size += sizeof(struct ibv_flow_spec_tcp_udp); - detected_items |= tunnel ? - MLX5_FLOW_LAYER_INNER_L4_TCP : - MLX5_FLOW_LAYER_OUTER_L4_TCP; break; case RTE_FLOW_ITEM_TYPE_VXLAN: size += sizeof(struct ibv_flow_spec_tunnel); - detected_items |= MLX5_FLOW_LAYER_VXLAN; break; case RTE_FLOW_ITEM_TYPE_VXLAN_GPE: size += sizeof(struct ibv_flow_spec_tunnel); - detected_items |= MLX5_FLOW_LAYER_VXLAN_GPE; break; #ifdef HAVE_IBV_DEVICE_MPLS_SUPPORT case RTE_FLOW_ITEM_TYPE_GRE: size += sizeof(struct ibv_flow_spec_gre); - detected_items |= MLX5_FLOW_LAYER_GRE; break; case RTE_FLOW_ITEM_TYPE_MPLS: size += sizeof(struct ibv_flow_spec_mpls); - detected_items |= MLX5_FLOW_LAYER_MPLS; break; #else case RTE_FLOW_ITEM_TYPE_GRE: size += sizeof(struct ibv_flow_spec_tunnel); - detected_items |= MLX5_FLOW_LAYER_TUNNEL; break; #endif default: break; } } - *item_flags = detected_items; return size; } @@ -1458,16 +1566,14 @@ flow_verbs_get_items_and_size(const struct rte_flow_item items[], * The required size is calculate based on the actions and items. This function * also returns the detected actions and items for later use. * + * @param[in] dev + * Pointer to Ethernet device. * @param[in] attr * Pointer to the flow attributes. * @param[in] items * Pointer to the list of items. * @param[in] actions * Pointer to the list of actions. - * @param[out] item_flags - * Pointer to bit mask of all items detected. - * @param[out] action_flags - * Pointer to bit mask of all actions detected. * @param[out] error * Pointer to the error structure. * @@ -1476,29 +1582,54 @@ flow_verbs_get_items_and_size(const struct rte_flow_item items[], * is set. */ static struct mlx5_flow * -flow_verbs_prepare(const struct rte_flow_attr *attr __rte_unused, +flow_verbs_prepare(struct rte_eth_dev *dev, + const struct rte_flow_attr *attr __rte_unused, const struct rte_flow_item items[], const struct rte_flow_action actions[], - uint64_t *item_flags, - uint64_t *action_flags, struct rte_flow_error *error) { - uint32_t size = sizeof(struct mlx5_flow) + sizeof(struct ibv_flow_attr); - struct mlx5_flow *flow; - - size += flow_verbs_get_actions_and_size(actions, action_flags); - size += flow_verbs_get_items_and_size(items, item_flags); - flow = rte_calloc(__func__, 1, size, 0); - if (!flow) { + size_t size = 0; + uint32_t handle_idx = 0; + struct mlx5_flow *dev_flow; + struct mlx5_flow_handle *dev_handle; + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_workspace *wks = mlx5_flow_get_thread_workspace(); + + MLX5_ASSERT(wks); + size += flow_verbs_get_actions_size(actions); + size += flow_verbs_get_items_size(items); + if (size > MLX5_VERBS_MAX_SPEC_ACT_SIZE) { + rte_flow_error_set(error, E2BIG, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, + "Verbs spec/action size too large"); + return NULL; + } + /* In case of corrupting the memory. */ + if (wks->flow_idx >= MLX5_NUM_MAX_DEV_FLOWS) { + rte_flow_error_set(error, ENOSPC, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, + "not free temporary device flow"); + return NULL; + } + dev_handle = mlx5_ipool_zmalloc(priv->sh->ipool[MLX5_IPOOL_MLX5_FLOW], + &handle_idx); + if (!dev_handle) { rte_flow_error_set(error, ENOMEM, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, - "not enough memory to create flow"); + "not enough memory to create flow handle"); return NULL; } - flow->verbs.attr = (void *)(flow + 1); - flow->verbs.specs = - (uint8_t *)(flow + 1) + sizeof(struct ibv_flow_attr); - return flow; + MLX5_ASSERT(wks->flow_idx + 1 < RTE_DIM(wks->flows)); + dev_flow = &wks->flows[wks->flow_idx++]; + dev_flow->handle = dev_handle; + dev_flow->handle_idx = handle_idx; + /* Memcpy is used, only size needs to be cleared to 0. */ + dev_flow->verbs.size = 0; + dev_flow->verbs.attr.num_of_specs = 0; + dev_flow->ingress = attr->ingress; + dev_flow->hash_fields = 0; + /* Need to set transfer attribute: not supported in Verbs mode. */ + return dev_flow; } /** @@ -1518,7 +1649,7 @@ flow_verbs_prepare(const struct rte_flow_attr *attr __rte_unused, * Pointer to the error structure. * * @return - * 0 on success, else a negative errno value otherwise and rte_ernno is set. + * 0 on success, else a negative errno value otherwise and rte_errno is set. */ static int flow_verbs_translate(struct rte_eth_dev *dev, @@ -1528,50 +1659,56 @@ flow_verbs_translate(struct rte_eth_dev *dev, const struct rte_flow_action actions[], struct rte_flow_error *error) { - uint64_t action_flags = 0; uint64_t item_flags = 0; + uint64_t action_flags = 0; uint64_t priority = attr->priority; - struct priv *priv = dev->data->dev_private; - - if (priority == MLX5_FLOW_PRIO_RSVD) - priority = priv->config.flow_prio - 1; + uint32_t subpriority = 0; + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_workspace *wks = mlx5_flow_get_thread_workspace(); + struct mlx5_flow_rss_desc *rss_desc; + + MLX5_ASSERT(wks); + rss_desc = &wks->rss_desc; + if (priority == MLX5_FLOW_LOWEST_PRIO_INDICATOR) + priority = priv->sh->flow_max_priority - 1; for (; actions->type != RTE_FLOW_ACTION_TYPE_END; actions++) { int ret; + switch (actions->type) { case RTE_FLOW_ACTION_TYPE_VOID: break; case RTE_FLOW_ACTION_TYPE_FLAG: - flow_verbs_translate_action_flag(actions, - &action_flags, - dev_flow); + flow_verbs_translate_action_flag(dev_flow, actions); + action_flags |= MLX5_FLOW_ACTION_FLAG; + dev_flow->handle->mark = 1; break; case RTE_FLOW_ACTION_TYPE_MARK: - flow_verbs_translate_action_mark(actions, - &action_flags, - dev_flow); + flow_verbs_translate_action_mark(dev_flow, actions); + action_flags |= MLX5_FLOW_ACTION_MARK; + dev_flow->handle->mark = 1; break; case RTE_FLOW_ACTION_TYPE_DROP: - flow_verbs_translate_action_drop(&action_flags, - dev_flow); + flow_verbs_translate_action_drop(dev_flow, actions); + action_flags |= MLX5_FLOW_ACTION_DROP; + dev_flow->handle->fate_action = MLX5_FLOW_FATE_DROP; break; case RTE_FLOW_ACTION_TYPE_QUEUE: - flow_verbs_translate_action_queue(actions, - &action_flags, - dev_flow); + flow_verbs_translate_action_queue(rss_desc, actions); + action_flags |= MLX5_FLOW_ACTION_QUEUE; + dev_flow->handle->fate_action = MLX5_FLOW_FATE_QUEUE; break; case RTE_FLOW_ACTION_TYPE_RSS: - flow_verbs_translate_action_rss(actions, - &action_flags, - dev_flow); + flow_verbs_translate_action_rss(rss_desc, actions); + action_flags |= MLX5_FLOW_ACTION_RSS; + dev_flow->handle->fate_action = MLX5_FLOW_FATE_QUEUE; break; case RTE_FLOW_ACTION_TYPE_COUNT: - ret = flow_verbs_translate_action_count(dev, + ret = flow_verbs_translate_action_count(dev_flow, actions, - &action_flags, - dev_flow, - error); + dev, error); if (ret < 0) return ret; + action_flags |= MLX5_FLOW_ACTION_COUNT; break; default: return rte_flow_error_set(error, ENOTSUP, @@ -1580,62 +1717,114 @@ flow_verbs_translate(struct rte_eth_dev *dev, "action not supported"); } } - /* Device flow should have action flags by flow_drv_prepare(). */ - assert(dev_flow->flow->actions == action_flags); + dev_flow->act_flags = action_flags; for (; items->type != RTE_FLOW_ITEM_TYPE_END; items++) { + int tunnel = !!(item_flags & MLX5_FLOW_LAYER_TUNNEL); + switch (items->type) { case RTE_FLOW_ITEM_TYPE_VOID: break; case RTE_FLOW_ITEM_TYPE_ETH: - flow_verbs_translate_item_eth(items, &item_flags, - dev_flow); + flow_verbs_translate_item_eth(dev_flow, items, + item_flags); + subpriority = MLX5_PRIORITY_MAP_L2; + item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L2 : + MLX5_FLOW_LAYER_OUTER_L2; break; case RTE_FLOW_ITEM_TYPE_VLAN: - flow_verbs_translate_item_vlan(items, &item_flags, - dev_flow); + flow_verbs_translate_item_vlan(dev_flow, items, + item_flags); + subpriority = MLX5_PRIORITY_MAP_L2; + item_flags |= tunnel ? (MLX5_FLOW_LAYER_INNER_L2 | + MLX5_FLOW_LAYER_INNER_VLAN) : + (MLX5_FLOW_LAYER_OUTER_L2 | + MLX5_FLOW_LAYER_OUTER_VLAN); break; case RTE_FLOW_ITEM_TYPE_IPV4: - flow_verbs_translate_item_ipv4(items, &item_flags, - dev_flow); + flow_verbs_translate_item_ipv4(dev_flow, items, + item_flags); + subpriority = MLX5_PRIORITY_MAP_L3; + dev_flow->hash_fields |= + mlx5_flow_hashfields_adjust + (rss_desc, tunnel, + MLX5_IPV4_LAYER_TYPES, + MLX5_IPV4_IBV_RX_HASH); + item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L3_IPV4 : + MLX5_FLOW_LAYER_OUTER_L3_IPV4; break; case RTE_FLOW_ITEM_TYPE_IPV6: - flow_verbs_translate_item_ipv6(items, &item_flags, - dev_flow); - break; - case RTE_FLOW_ITEM_TYPE_UDP: - flow_verbs_translate_item_udp(items, &item_flags, - dev_flow); + flow_verbs_translate_item_ipv6(dev_flow, items, + item_flags); + subpriority = MLX5_PRIORITY_MAP_L3; + dev_flow->hash_fields |= + mlx5_flow_hashfields_adjust + (rss_desc, tunnel, + MLX5_IPV6_LAYER_TYPES, + MLX5_IPV6_IBV_RX_HASH); + item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L3_IPV6 : + MLX5_FLOW_LAYER_OUTER_L3_IPV6; break; case RTE_FLOW_ITEM_TYPE_TCP: - flow_verbs_translate_item_tcp(items, &item_flags, - dev_flow); + flow_verbs_translate_item_tcp(dev_flow, items, + item_flags); + subpriority = MLX5_PRIORITY_MAP_L4; + if (dev_flow->hash_fields != 0) + dev_flow->hash_fields |= + mlx5_flow_hashfields_adjust + (rss_desc, tunnel, RTE_ETH_RSS_TCP, + (IBV_RX_HASH_SRC_PORT_TCP | + IBV_RX_HASH_DST_PORT_TCP)); + item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L4_TCP : + MLX5_FLOW_LAYER_OUTER_L4_TCP; + break; + case RTE_FLOW_ITEM_TYPE_UDP: + flow_verbs_translate_item_udp(dev_flow, items, + item_flags); + subpriority = MLX5_PRIORITY_MAP_L4; + if (dev_flow->hash_fields != 0) + dev_flow->hash_fields |= + mlx5_flow_hashfields_adjust + (rss_desc, tunnel, RTE_ETH_RSS_UDP, + (IBV_RX_HASH_SRC_PORT_UDP | + IBV_RX_HASH_DST_PORT_UDP)); + item_flags |= tunnel ? MLX5_FLOW_LAYER_INNER_L4_UDP : + MLX5_FLOW_LAYER_OUTER_L4_UDP; break; case RTE_FLOW_ITEM_TYPE_VXLAN: - flow_verbs_translate_item_vxlan(items, &item_flags, - dev_flow); + flow_verbs_translate_item_vxlan(dev_flow, items, + item_flags); + subpriority = MLX5_TUNNEL_PRIO_GET(rss_desc); + item_flags |= MLX5_FLOW_LAYER_VXLAN; break; case RTE_FLOW_ITEM_TYPE_VXLAN_GPE: - flow_verbs_translate_item_vxlan_gpe(items, &item_flags, - dev_flow); + flow_verbs_translate_item_vxlan_gpe(dev_flow, items, + item_flags); + subpriority = MLX5_TUNNEL_PRIO_GET(rss_desc); + item_flags |= MLX5_FLOW_LAYER_VXLAN_GPE; break; case RTE_FLOW_ITEM_TYPE_GRE: - flow_verbs_translate_item_gre(items, &item_flags, - dev_flow); + flow_verbs_translate_item_gre(dev_flow, items, + item_flags); + subpriority = MLX5_TUNNEL_PRIO_GET(rss_desc); + item_flags |= MLX5_FLOW_LAYER_GRE; break; case RTE_FLOW_ITEM_TYPE_MPLS: - flow_verbs_translate_item_mpls(items, &item_flags, - dev_flow); + flow_verbs_translate_item_mpls(dev_flow, items, + item_flags); + subpriority = MLX5_TUNNEL_PRIO_GET(rss_desc); + item_flags |= MLX5_FLOW_LAYER_MPLS; break; default: return rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM, - NULL, - "item not supported"); + NULL, "item not supported"); } } - dev_flow->verbs.attr->priority = - mlx5_flow_adjust_priority(dev, priority, - dev_flow->verbs.attr->priority); + dev_flow->handle->layers = item_flags; + /* Other members of attr will be ignored. */ + dev_flow->verbs.attr.priority = + mlx5_flow_adjust_priority(dev, priority, subpriority); + dev_flow->verbs.attr.port = (uint8_t)priv->dev_port; return 0; } @@ -1650,28 +1839,26 @@ flow_verbs_translate(struct rte_eth_dev *dev, static void flow_verbs_remove(struct rte_eth_dev *dev, struct rte_flow *flow) { - struct mlx5_flow_verbs *verbs; - struct mlx5_flow *dev_flow; + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_handle *handle; + uint32_t handle_idx; if (!flow) return; - LIST_FOREACH(dev_flow, &flow->dev_flows, next) { - verbs = &dev_flow->verbs; - if (verbs->flow) { - claim_zero(mlx5_glue->destroy_flow(verbs->flow)); - verbs->flow = NULL; + SILIST_FOREACH(priv->sh->ipool[MLX5_IPOOL_MLX5_FLOW], flow->dev_handles, + handle_idx, handle, next) { + if (handle->drv_flow) { + claim_zero(mlx5_glue->destroy_flow(handle->drv_flow)); + handle->drv_flow = NULL; } - if (verbs->hrxq) { - if (flow->actions & MLX5_FLOW_ACTION_DROP) - mlx5_hrxq_drop_release(dev); - else - mlx5_hrxq_release(dev, verbs->hrxq); - verbs->hrxq = NULL; + /* hrxq is union, don't touch it only the flag is set. */ + if (handle->rix_hrxq && + handle->fate_action == MLX5_FLOW_FATE_QUEUE) { + mlx5_hrxq_release(dev, handle->rix_hrxq); + handle->rix_hrxq = 0; } - } - if (flow->counter) { - flow_verbs_counter_release(flow->counter); - flow->counter = NULL; + if (handle->vf_vlan.tag && handle->vf_vlan.created) + mlx5_vlan_vmwa_release(dev, &handle->vf_vlan); } } @@ -1686,15 +1873,26 @@ flow_verbs_remove(struct rte_eth_dev *dev, struct rte_flow *flow) static void flow_verbs_destroy(struct rte_eth_dev *dev, struct rte_flow *flow) { - struct mlx5_flow *dev_flow; + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_handle *handle; if (!flow) return; flow_verbs_remove(dev, flow); - while (!LIST_EMPTY(&flow->dev_flows)) { - dev_flow = LIST_FIRST(&flow->dev_flows); - LIST_REMOVE(dev_flow, next); - rte_free(dev_flow); + while (flow->dev_handles) { + uint32_t tmp_idx = flow->dev_handles; + + handle = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_MLX5_FLOW], + tmp_idx); + if (!handle) + return; + flow->dev_handles = handle->next.next; + mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_MLX5_FLOW], + tmp_idx); + } + if (flow->counter) { + flow_verbs_counter_release(dev, flow->counter); + flow->counter = 0; } } @@ -1715,37 +1913,35 @@ static int flow_verbs_apply(struct rte_eth_dev *dev, struct rte_flow *flow, struct rte_flow_error *error) { - struct mlx5_flow_verbs *verbs; + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_handle *handle; struct mlx5_flow *dev_flow; + struct mlx5_hrxq *hrxq; + uint32_t dev_handles; int err; - - LIST_FOREACH(dev_flow, &flow->dev_flows, next) { - verbs = &dev_flow->verbs; - if (flow->actions & MLX5_FLOW_ACTION_DROP) { - verbs->hrxq = mlx5_hrxq_drop_new(dev); - if (!verbs->hrxq) { - rte_flow_error_set - (error, errno, - RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, - "cannot get drop hash queue"); - goto error; - } + int idx; + struct mlx5_flow_workspace *wks = mlx5_flow_get_thread_workspace(); + + MLX5_ASSERT(wks); + for (idx = wks->flow_idx - 1; idx >= 0; idx--) { + dev_flow = &wks->flows[idx]; + handle = dev_flow->handle; + if (handle->fate_action == MLX5_FLOW_FATE_DROP) { + MLX5_ASSERT(priv->drop_queue.hrxq); + hrxq = priv->drop_queue.hrxq; } else { - struct mlx5_hrxq *hrxq; - - hrxq = mlx5_hrxq_get(dev, flow->key, - MLX5_RSS_HASH_KEY_LEN, - verbs->hash_fields, - (*flow->queue), - flow->rss.queue_num); - if (!hrxq) - hrxq = mlx5_hrxq_new(dev, flow->key, - MLX5_RSS_HASH_KEY_LEN, - verbs->hash_fields, - (*flow->queue), - flow->rss.queue_num, - !!(dev_flow->layers & - MLX5_FLOW_LAYER_TUNNEL)); + uint32_t hrxq_idx; + struct mlx5_flow_rss_desc *rss_desc = &wks->rss_desc; + + MLX5_ASSERT(rss_desc->queue_num); + rss_desc->key_len = MLX5_RSS_HASH_KEY_LEN; + rss_desc->hash_fields = dev_flow->hash_fields; + rss_desc->tunnel = !!(handle->layers & + MLX5_FLOW_LAYER_TUNNEL); + rss_desc->shared_rss = 0; + hrxq_idx = mlx5_hrxq_get(dev, rss_desc); + hrxq = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_HRXQ], + hrxq_idx); if (!hrxq) { rte_flow_error_set (error, rte_errno, @@ -1753,30 +1949,42 @@ flow_verbs_apply(struct rte_eth_dev *dev, struct rte_flow *flow, "cannot get hash queue"); goto error; } - verbs->hrxq = hrxq; + handle->rix_hrxq = hrxq_idx; } - verbs->flow = mlx5_glue->create_flow(verbs->hrxq->qp, - verbs->attr); - if (!verbs->flow) { + MLX5_ASSERT(hrxq); + handle->drv_flow = mlx5_glue->create_flow + (hrxq->qp, &dev_flow->verbs.attr); + if (!handle->drv_flow) { rte_flow_error_set(error, errno, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, "hardware refuses to create flow"); goto error; } + if (priv->vmwa_context && + handle->vf_vlan.tag && !handle->vf_vlan.created) { + /* + * The rule contains the VLAN pattern. + * For VF we are going to create VLAN + * interface to make hypervisor set correct + * e-Switch vport context. + */ + mlx5_vlan_vmwa_acquire(dev, &handle->vf_vlan); + } } return 0; error: err = rte_errno; /* Save rte_errno before cleanup. */ - LIST_FOREACH(dev_flow, &flow->dev_flows, next) { - verbs = &dev_flow->verbs; - if (verbs->hrxq) { - if (flow->actions & MLX5_FLOW_ACTION_DROP) - mlx5_hrxq_drop_release(dev); - else - mlx5_hrxq_release(dev, verbs->hrxq); - verbs->hrxq = NULL; + SILIST_FOREACH(priv->sh->ipool[MLX5_IPOOL_MLX5_FLOW], flow->dev_handles, + dev_handles, handle, next) { + /* hrxq is union, don't touch it only the flag is set. */ + if (handle->rix_hrxq && + handle->fate_action == MLX5_FLOW_FATE_QUEUE) { + mlx5_hrxq_release(dev, handle->rix_hrxq); + handle->rix_hrxq = 0; } + if (handle->vf_vlan.tag && handle->vf_vlan.created) + mlx5_vlan_vmwa_release(dev, &handle->vf_vlan); } rte_errno = err; /* Restore rte_errno. */ return -rte_errno; @@ -1814,6 +2022,17 @@ flow_verbs_query(struct rte_eth_dev *dev, return ret; } +static int +flow_verbs_sync_domain(struct rte_eth_dev *dev, uint32_t domains, + uint32_t flags) +{ + RTE_SET_USED(dev); + RTE_SET_USED(domains); + RTE_SET_USED(flags); + + return 0; +} + const struct mlx5_flow_driver_ops mlx5_flow_verbs_drv_ops = { .validate = flow_verbs_validate, .prepare = flow_verbs_prepare, @@ -1822,4 +2041,6 @@ const struct mlx5_flow_driver_ops mlx5_flow_verbs_drv_ops = { .remove = flow_verbs_remove, .destroy = flow_verbs_destroy, .query = flow_verbs_query, + .sync_domain = flow_verbs_sync_domain, + .discover_priorities = flow_verbs_discover_priorities, };