X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fmlx5%2Fmlx5_flow_dv.c;h=8b018fb3cf2da010dd4d7d1c19c84f87846fd0f7;hb=4c204fe5e5d2bf0001ab0d017df9f0f578b0434c;hp=d7176c80efcb8dafc002c86bd08da6efc8dfc90d;hpb=82e4aca2caa011c85911615dc38e17da86c5a184;p=dpdk.git diff --git a/drivers/net/mlx5/mlx5_flow_dv.c b/drivers/net/mlx5/mlx5_flow_dv.c index d7176c80ef..8b018fb3cf 100644 --- a/drivers/net/mlx5/mlx5_flow_dv.c +++ b/drivers/net/mlx5/mlx5_flow_dv.c @@ -24,16 +24,19 @@ #include #include #include +#include #include #include #include #include -#include "mlx5.h" +#include +#include +#include + #include "mlx5_defs.h" -#include "mlx5_glue.h" +#include "mlx5.h" #include "mlx5_flow.h" -#include "mlx5_prm.h" #include "mlx5_rxtx.h" #ifdef HAVE_IBV_FLOW_DV_SUPPORT @@ -52,8 +55,6 @@ #define MLX5DV_DR_ACTION_FLAGS_ROOT_LEVEL 1 #endif -#define MLX5_ENCAPSULATION_DECISION_SIZE (sizeof(struct rte_flow_item_eth) + \ - sizeof(struct rte_flow_item_ipv4)) /* VLAN header definitions */ #define MLX5DV_FLOW_VLAN_PCP_SHIFT 13 #define MLX5DV_FLOW_VLAN_PCP_MASK (0x7 << MLX5DV_FLOW_VLAN_PCP_SHIFT) @@ -73,6 +74,10 @@ union flow_dv_attr { uint32_t attr; }; +static int +flow_dv_tbl_resource_release(struct rte_eth_dev *dev, + struct mlx5_flow_tbl_resource *tbl); + /** * Initialize flow attributes structure according to flow items' types. * @@ -83,19 +88,76 @@ union flow_dv_attr { * Pointer to item specification. * @param[out] attr * Pointer to flow attributes structure. + * @param[in] dev_flow + * Pointer to the sub flow. + * @param[in] tunnel_decap + * Whether action is after tunnel decapsulation. */ static void -flow_dv_attr_init(const struct rte_flow_item *item, union flow_dv_attr *attr) +flow_dv_attr_init(const struct rte_flow_item *item, union flow_dv_attr *attr, + struct mlx5_flow *dev_flow, bool tunnel_decap) { + uint64_t layers = dev_flow->handle->layers; + + /* + * If layers is already initialized, it means this dev_flow is the + * suffix flow, the layers flags is set by the prefix flow. Need to + * use the layer flags from prefix flow as the suffix flow may not + * have the user defined items as the flow is split. + */ + if (layers) { + if (layers & MLX5_FLOW_LAYER_OUTER_L3_IPV4) + attr->ipv4 = 1; + else if (layers & MLX5_FLOW_LAYER_OUTER_L3_IPV6) + attr->ipv6 = 1; + if (layers & MLX5_FLOW_LAYER_OUTER_L4_TCP) + attr->tcp = 1; + else if (layers & MLX5_FLOW_LAYER_OUTER_L4_UDP) + attr->udp = 1; + attr->valid = 1; + return; + } for (; item->type != RTE_FLOW_ITEM_TYPE_END; item++) { + uint8_t next_protocol = 0xff; switch (item->type) { + case RTE_FLOW_ITEM_TYPE_GRE: + case RTE_FLOW_ITEM_TYPE_NVGRE: + case RTE_FLOW_ITEM_TYPE_VXLAN: + case RTE_FLOW_ITEM_TYPE_VXLAN_GPE: + case RTE_FLOW_ITEM_TYPE_GENEVE: + case RTE_FLOW_ITEM_TYPE_MPLS: + if (tunnel_decap) + attr->attr = 0; + break; case RTE_FLOW_ITEM_TYPE_IPV4: if (!attr->ipv6) attr->ipv4 = 1; + if (item->mask != NULL && + ((const struct rte_flow_item_ipv4 *) + item->mask)->hdr.next_proto_id) + next_protocol = + ((const struct rte_flow_item_ipv4 *) + (item->spec))->hdr.next_proto_id & + ((const struct rte_flow_item_ipv4 *) + (item->mask))->hdr.next_proto_id; + if ((next_protocol == IPPROTO_IPIP || + next_protocol == IPPROTO_IPV6) && tunnel_decap) + attr->attr = 0; break; case RTE_FLOW_ITEM_TYPE_IPV6: if (!attr->ipv4) attr->ipv6 = 1; + if (item->mask != NULL && + ((const struct rte_flow_item_ipv6 *) + item->mask)->hdr.proto) + next_protocol = + ((const struct rte_flow_item_ipv6 *) + (item->spec))->hdr.proto & + ((const struct rte_flow_item_ipv6 *) + (item->mask))->hdr.proto; + if ((next_protocol == IPPROTO_IPIP || + next_protocol == IPPROTO_IPV6) && tunnel_decap) + attr->attr = 0; break; case RTE_FLOW_ITEM_TYPE_UDP: if (!attr->tcp) @@ -198,8 +260,8 @@ mlx5_flow_tunnel_ip_check(const struct rte_flow_item *item __rte_unused, uint8_t next_protocol, uint64_t *item_flags, int *tunnel) { - assert(item->type == RTE_FLOW_ITEM_TYPE_IPV4 || - item->type == RTE_FLOW_ITEM_TYPE_IPV6); + MLX5_ASSERT(item->type == RTE_FLOW_ITEM_TYPE_IPV4 || + item->type == RTE_FLOW_ITEM_TYPE_IPV6); if (next_protocol == IPPROTO_IPIP) { *item_flags |= MLX5_FLOW_LAYER_IPIP; *tunnel = 1; @@ -229,7 +291,7 @@ flow_dv_shared_lock(struct rte_eth_dev *dev) int ret; ret = pthread_mutex_lock(&sh->dv_mutex); - assert(!ret); + MLX5_ASSERT(!ret); (void)ret; } } @@ -244,7 +306,7 @@ flow_dv_shared_unlock(struct rte_eth_dev *dev) int ret; ret = pthread_mutex_unlock(&sh->dv_mutex); - assert(!ret); + MLX5_ASSERT(!ret); (void)ret; } } @@ -308,7 +370,7 @@ flow_dv_fetch_field(const uint8_t *data, uint32_t size) ret = rte_be_to_cpu_32(*(const unaligned_uint32_t *)data); break; default: - assert(false); + MLX5_ASSERT(false); ret = 0; break; } @@ -358,8 +420,8 @@ flow_dv_convert_modify_action(struct rte_flow_item *item, * The fields should be presented as in big-endian format either. * Mask must be always present, it defines the actual field width. */ - assert(item->mask); - assert(field->size); + MLX5_ASSERT(item->mask); + MLX5_ASSERT(field->size); do { unsigned int size_b; unsigned int off_b; @@ -381,23 +443,25 @@ flow_dv_convert_modify_action(struct rte_flow_item *item, off_b = rte_bsf32(mask); size_b = sizeof(uint32_t) * CHAR_BIT - off_b - __builtin_clz(mask); - assert(size_b); + MLX5_ASSERT(size_b); size_b = size_b == sizeof(uint32_t) * CHAR_BIT ? 0 : size_b; - actions[i].action_type = type; - actions[i].field = field->id; - actions[i].offset = off_b; - actions[i].length = size_b; + actions[i] = (struct mlx5_modification_cmd) { + .action_type = type, + .field = field->id, + .offset = off_b, + .length = size_b, + }; /* Convert entire record to expected big-endian format. */ actions[i].data0 = rte_cpu_to_be_32(actions[i].data0); if (type == MLX5_MODIFICATION_TYPE_COPY) { - assert(dcopy); + MLX5_ASSERT(dcopy); actions[i].dst_field = dcopy->id; actions[i].dst_offset = (int)dcopy->offset < 0 ? off_b : dcopy->offset; /* Convert entire record to big-endian format. */ actions[i].data1 = rte_cpu_to_be_32(actions[i].data1); } else { - assert(item->spec); + MLX5_ASSERT(item->spec); data = flow_dv_fetch_field((const uint8_t *)item->spec + field->offset, field->size); /* Shift out the trailing masked bits from data. */ @@ -569,17 +633,19 @@ flow_dv_convert_action_modify_vlan_vid const struct rte_flow_action_of_set_vlan_vid *conf = (const struct rte_flow_action_of_set_vlan_vid *)(action->conf); int i = resource->actions_num; - struct mlx5_modification_cmd *actions = &resource->actions[i]; + struct mlx5_modification_cmd *actions = resource->actions; struct field_modify_info *field = modify_vlan_out_first_vid; if (i >= MLX5_MAX_MODIFY_NUM) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, "too many items to modify"); - actions[i].action_type = MLX5_MODIFICATION_TYPE_SET; - actions[i].field = field->id; - actions[i].length = field->size; - actions[i].offset = field->offset; + actions[i] = (struct mlx5_modification_cmd) { + .action_type = MLX5_MODIFICATION_TYPE_SET, + .field = field->id, + .length = field->size, + .offset = field->offset, + }; actions[i].data0 = rte_cpu_to_be_32(actions[i].data0); actions[i].data1 = conf->vlan_vid; actions[i].data1 = actions[i].data1 << 16; @@ -598,6 +664,10 @@ flow_dv_convert_action_modify_vlan_vid * Pointer to rte_flow_item objects list. * @param[in] attr * Pointer to flow attributes structure. + * @param[in] dev_flow + * Pointer to the sub flow. + * @param[in] tunnel_decap + * Whether action is after tunnel decapsulation. * @param[out] error * Pointer to the error structure. * @@ -609,8 +679,8 @@ flow_dv_convert_action_modify_tp (struct mlx5_flow_dv_modify_hdr_resource *resource, const struct rte_flow_action *action, const struct rte_flow_item *items, - union flow_dv_attr *attr, - struct rte_flow_error *error) + union flow_dv_attr *attr, struct mlx5_flow *dev_flow, + bool tunnel_decap, struct rte_flow_error *error) { const struct rte_flow_action_set_tp *conf = (const struct rte_flow_action_set_tp *)(action->conf); @@ -622,7 +692,7 @@ flow_dv_convert_action_modify_tp struct field_modify_info *field; if (!attr->valid) - flow_dv_attr_init(items, attr); + flow_dv_attr_init(items, attr, dev_flow, tunnel_decap); if (attr->udp) { memset(&udp, 0, sizeof(udp)); memset(&udp_mask, 0, sizeof(udp_mask)); @@ -639,8 +709,8 @@ flow_dv_convert_action_modify_tp item.spec = &udp; item.mask = &udp_mask; field = modify_udp; - } - if (attr->tcp) { + } else { + MLX5_ASSERT(attr->tcp); memset(&tcp, 0, sizeof(tcp)); memset(&tcp_mask, 0, sizeof(tcp_mask)); if (action->type == RTE_FLOW_ACTION_TYPE_SET_TP_SRC) { @@ -672,6 +742,10 @@ flow_dv_convert_action_modify_tp * Pointer to rte_flow_item objects list. * @param[in] attr * Pointer to flow attributes structure. + * @param[in] dev_flow + * Pointer to the sub flow. + * @param[in] tunnel_decap + * Whether action is after tunnel decapsulation. * @param[out] error * Pointer to the error structure. * @@ -683,8 +757,8 @@ flow_dv_convert_action_modify_ttl (struct mlx5_flow_dv_modify_hdr_resource *resource, const struct rte_flow_action *action, const struct rte_flow_item *items, - union flow_dv_attr *attr, - struct rte_flow_error *error) + union flow_dv_attr *attr, struct mlx5_flow *dev_flow, + bool tunnel_decap, struct rte_flow_error *error) { const struct rte_flow_action_set_ttl *conf = (const struct rte_flow_action_set_ttl *)(action->conf); @@ -696,7 +770,7 @@ flow_dv_convert_action_modify_ttl struct field_modify_info *field; if (!attr->valid) - flow_dv_attr_init(items, attr); + flow_dv_attr_init(items, attr, dev_flow, tunnel_decap); if (attr->ipv4) { memset(&ipv4, 0, sizeof(ipv4)); memset(&ipv4_mask, 0, sizeof(ipv4_mask)); @@ -706,8 +780,8 @@ flow_dv_convert_action_modify_ttl item.spec = &ipv4; item.mask = &ipv4_mask; field = modify_ipv4; - } - if (attr->ipv6) { + } else { + MLX5_ASSERT(attr->ipv6); memset(&ipv6, 0, sizeof(ipv6)); memset(&ipv6_mask, 0, sizeof(ipv6_mask)); ipv6.hdr.hop_limits = conf->ttl_value; @@ -732,6 +806,10 @@ flow_dv_convert_action_modify_ttl * Pointer to rte_flow_item objects list. * @param[in] attr * Pointer to flow attributes structure. + * @param[in] dev_flow + * Pointer to the sub flow. + * @param[in] tunnel_decap + * Whether action is after tunnel decapsulation. * @param[out] error * Pointer to the error structure. * @@ -742,8 +820,8 @@ static int flow_dv_convert_action_modify_dec_ttl (struct mlx5_flow_dv_modify_hdr_resource *resource, const struct rte_flow_item *items, - union flow_dv_attr *attr, - struct rte_flow_error *error) + union flow_dv_attr *attr, struct mlx5_flow *dev_flow, + bool tunnel_decap, struct rte_flow_error *error) { struct rte_flow_item item; struct rte_flow_item_ipv4 ipv4; @@ -753,7 +831,7 @@ flow_dv_convert_action_modify_dec_ttl struct field_modify_info *field; if (!attr->valid) - flow_dv_attr_init(items, attr); + flow_dv_attr_init(items, attr, dev_flow, tunnel_decap); if (attr->ipv4) { memset(&ipv4, 0, sizeof(ipv4)); memset(&ipv4_mask, 0, sizeof(ipv4_mask)); @@ -763,8 +841,8 @@ flow_dv_convert_action_modify_dec_ttl item.spec = &ipv4; item.mask = &ipv4_mask; field = modify_ipv4; - } - if (attr->ipv6) { + } else { + MLX5_ASSERT(attr->ipv6); memset(&ipv6, 0, sizeof(ipv6)); memset(&ipv6_mask, 0, sizeof(ipv6_mask)); ipv6.hdr.hop_limits = 0xFF; @@ -909,10 +987,12 @@ flow_dv_convert_action_set_reg return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, "too many items to modify"); - assert(conf->id != REG_NONE); - assert(conf->id < RTE_DIM(reg_to_field)); - actions[i].action_type = MLX5_MODIFICATION_TYPE_SET; - actions[i].field = reg_to_field[conf->id]; + MLX5_ASSERT(conf->id != REG_NONE); + MLX5_ASSERT(conf->id < RTE_DIM(reg_to_field)); + actions[i] = (struct mlx5_modification_cmd) { + .action_type = MLX5_MODIFICATION_TYPE_SET, + .field = reg_to_field[conf->id], + }; actions[i].data0 = rte_cpu_to_be_32(actions[i].data0); actions[i].data1 = rte_cpu_to_be_32(conf->data); ++i; @@ -957,10 +1037,10 @@ flow_dv_convert_action_set_tag ret = mlx5_flow_get_reg_id(dev, MLX5_APP_TAG, conf->index, error); if (ret < 0) return ret; - assert(ret != REG_NONE); - assert((unsigned int)ret < RTE_DIM(reg_to_field)); + MLX5_ASSERT(ret != REG_NONE); + MLX5_ASSERT((unsigned int)ret < RTE_DIM(reg_to_field)); reg_type = reg_to_field[ret]; - assert(reg_type > 0); + MLX5_ASSERT(reg_type > 0); reg_c_x[0] = (struct field_modify_info){4, 0, reg_type}; return flow_dv_convert_modify_action(&item, reg_c_x, NULL, resource, MLX5_MODIFICATION_TYPE_SET, error); @@ -1006,8 +1086,8 @@ flow_dv_convert_action_copy_mreg(struct rte_eth_dev *dev, struct mlx5_priv *priv = dev->data->dev_private; uint32_t reg_c0 = priv->sh->dv_regc0_mask; - assert(reg_c0); - assert(priv->config.dv_xmeta_en != MLX5_XMETA_MODE_LEGACY); + MLX5_ASSERT(reg_c0); + MLX5_ASSERT(priv->config.dv_xmeta_en != MLX5_XMETA_MODE_LEGACY); if (conf->dst == REG_C_0) { /* Copy to reg_c[0], within mask only. */ reg_dst.offset = rte_bsf32(reg_c0); @@ -1077,7 +1157,7 @@ flow_dv_convert_action_mark(struct rte_eth_dev *dev, {4, 0, 0}, /* dynamic instead of MLX5_MODI_META_REG_C_1. */ {0, 0, 0}, }; - enum modify_reg reg; + int reg; if (!mask) return rte_flow_error_set(error, EINVAL, @@ -1086,7 +1166,7 @@ flow_dv_convert_action_mark(struct rte_eth_dev *dev, reg = mlx5_flow_get_reg_id(dev, MLX5_FLOW_MARK, 0, error); if (reg < 0) return reg; - assert(reg > 0); + MLX5_ASSERT(reg > 0); if (reg == REG_C_0) { uint32_t msk_c0 = priv->sh->dv_regc0_mask; uint32_t shl_c0 = rte_bsf32(msk_c0); @@ -1119,7 +1199,7 @@ flow_dv_get_metadata_reg(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, struct rte_flow_error *error) { - enum modify_reg reg = + int reg = mlx5_flow_get_reg_id(dev, attr->transfer ? MLX5_METADATA_FDB : attr->egress ? @@ -1167,7 +1247,7 @@ flow_dv_convert_action_set_meta struct field_modify_info reg_c_x[] = { [1] = {0, 0, 0}, }; - enum modify_reg reg = flow_dv_get_metadata_reg(dev, attr, error); + int reg = flow_dv_get_metadata_reg(dev, attr, error); if (reg < 0) return reg; @@ -1181,7 +1261,7 @@ flow_dv_convert_action_set_meta uint32_t msk_c0 = priv->sh->dv_regc0_mask; uint32_t shl_c0; - assert(msk_c0); + MLX5_ASSERT(msk_c0); #if RTE_BYTE_ORDER == RTE_BIG_ENDIAN shl_c0 = rte_bsf32(msk_c0); #else @@ -1189,7 +1269,7 @@ flow_dv_convert_action_set_meta #endif mask <<= shl_c0; data <<= shl_c0; - assert(!(~msk_c0 & rte_cpu_to_be_32(mask))); + MLX5_ASSERT(!(~msk_c0 & rte_cpu_to_be_32(mask))); } reg_c_x[0] = (struct field_modify_info){4, 0, reg_to_field[reg]}; /* The routine expects parameters in memory as big-endian ones. */ @@ -1333,6 +1413,11 @@ flow_dv_validate_item_mark(struct rte_eth_dev *dev, "mark id exceeds the limit"); if (!mask) mask = &nic_mask; + if (!mask->id) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ITEM_SPEC, NULL, + "mask cannot be zero"); + ret = mlx5_flow_item_acceptable(item, (const uint8_t *)mask, (const uint8_t *)&nic_mask, sizeof(struct rte_flow_item_mark), @@ -1370,7 +1455,7 @@ flow_dv_validate_item_meta(struct rte_eth_dev *dev __rte_unused, struct rte_flow_item_meta nic_mask = { .data = UINT32_MAX }; - enum modify_reg reg; + int reg; int ret; if (!spec) @@ -1378,10 +1463,6 @@ flow_dv_validate_item_meta(struct rte_eth_dev *dev __rte_unused, RTE_FLOW_ERROR_TYPE_ITEM_SPEC, item->spec, "data cannot be empty"); - if (!spec->data) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ITEM_SPEC, NULL, - "data cannot be zero"); if (config->dv_xmeta_en != MLX5_XMETA_MODE_LEGACY) { if (!mlx5_flow_ext_mreg_supported(dev)) return rte_flow_error_set(error, ENOTSUP, @@ -1401,6 +1482,11 @@ flow_dv_validate_item_meta(struct rte_eth_dev *dev __rte_unused, } if (!mask) mask = &rte_flow_item_meta_mask; + if (!mask->data) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ITEM_SPEC, NULL, + "mask cannot be zero"); + ret = mlx5_flow_item_acceptable(item, (const uint8_t *)mask, (const uint8_t *)&nic_mask, sizeof(struct rte_flow_item_meta), @@ -1449,6 +1535,11 @@ flow_dv_validate_item_tag(struct rte_eth_dev *dev, "data cannot be empty"); if (!mask) mask = &rte_flow_item_tag_mask; + if (!mask->data) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ITEM_SPEC, NULL, + "mask cannot be zero"); + ret = mlx5_flow_item_acceptable(item, (const uint8_t *)mask, (const uint8_t *)&nic_mask, sizeof(struct rte_flow_item_tag), @@ -1463,7 +1554,7 @@ flow_dv_validate_item_tag(struct rte_eth_dev *dev, ret = mlx5_flow_get_reg_id(dev, MLX5_APP_TAG, spec->index, error); if (ret < 0) return ret; - assert(ret != REG_NONE); + MLX5_ASSERT(ret != REG_NONE); return 0; } @@ -1548,6 +1639,91 @@ flow_dv_validate_item_port_id(struct rte_eth_dev *dev, return 0; } +/* + * GTP flags are contained in 1 byte of the format: + * ------------------------------------------- + * | bit | 0 - 2 | 3 | 4 | 5 | 6 | 7 | + * |-----------------------------------------| + * | value | Version | PT | Res | E | S | PN | + * ------------------------------------------- + * + * Matching is supported only for GTP flags E, S, PN. + */ +#define MLX5_GTP_FLAGS_MASK 0x07 + +/** + * Validate VLAN item. + * + * @param[in] item + * Item specification. + * @param[in] item_flags + * Bit-fields that holds the items detected until now. + * @param[in] dev + * Ethernet device flow is being created on. + * @param[out] error + * Pointer to error structure. + * + * @return + * 0 on success, a negative errno value otherwise and rte_errno is set. + */ +static int +flow_dv_validate_item_vlan(const struct rte_flow_item *item, + uint64_t item_flags, + struct rte_eth_dev *dev, + struct rte_flow_error *error) +{ + const struct rte_flow_item_vlan *mask = item->mask; + const struct rte_flow_item_vlan nic_mask = { + .tci = RTE_BE16(UINT16_MAX), + .inner_type = RTE_BE16(UINT16_MAX), + }; + const int tunnel = !!(item_flags & MLX5_FLOW_LAYER_TUNNEL); + int ret; + const uint64_t l34m = tunnel ? (MLX5_FLOW_LAYER_INNER_L3 | + MLX5_FLOW_LAYER_INNER_L4) : + (MLX5_FLOW_LAYER_OUTER_L3 | + MLX5_FLOW_LAYER_OUTER_L4); + const uint64_t vlanm = tunnel ? MLX5_FLOW_LAYER_INNER_VLAN : + MLX5_FLOW_LAYER_OUTER_VLAN; + + if (item_flags & vlanm) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ITEM, item, + "multiple VLAN layers not supported"); + else if ((item_flags & l34m) != 0) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ITEM, item, + "VLAN cannot follow L3/L4 layer"); + if (!mask) + mask = &rte_flow_item_vlan_mask; + ret = mlx5_flow_item_acceptable(item, (const uint8_t *)mask, + (const uint8_t *)&nic_mask, + sizeof(struct rte_flow_item_vlan), + error); + if (ret) + return ret; + if (!tunnel && mask->tci != RTE_BE16(0x0fff)) { + struct mlx5_priv *priv = dev->data->dev_private; + + if (priv->vmwa_context) { + /* + * Non-NULL context means we have a virtual machine + * and SR-IOV enabled, we have to create VLAN interface + * to make hypervisor to setup E-Switch vport + * context correctly. We avoid creating the multiple + * VLAN interfaces, so we cannot support VLAN tag mask. + */ + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ITEM, + item, + "VLAN tag mask is not" + " supported in virtual" + " environment"); + } + } + return 0; +} + /** * Validate GTP item. * @@ -1570,8 +1746,10 @@ flow_dv_validate_item_gtp(struct rte_eth_dev *dev, struct rte_flow_error *error) { struct mlx5_priv *priv = dev->data->dev_private; + const struct rte_flow_item_gtp *spec = item->spec; const struct rte_flow_item_gtp *mask = item->mask; const struct rte_flow_item_gtp nic_mask = { + .v_pt_rsv_flags = MLX5_GTP_FLAGS_MASK, .msg_type = 0xff, .teid = RTE_BE32(0xffffffff), }; @@ -1591,6 +1769,11 @@ flow_dv_validate_item_gtp(struct rte_eth_dev *dev, "no outer UDP layer found"); if (!mask) mask = &rte_flow_item_gtp_mask; + if (spec && spec->v_pt_rsv_flags & ~MLX5_GTP_FLAGS_MASK) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ITEM, item, + "Match is supported for GTP" + " flags only"); return mlx5_flow_item_acceptable (item, (const uint8_t *)mask, (const uint8_t *)&nic_mask, @@ -1625,7 +1808,7 @@ flow_dv_validate_action_pop_vlan(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, struct rte_flow_error *error) { - struct mlx5_priv *priv = dev->data->dev_private; + const struct mlx5_priv *priv = dev->data->dev_private; (void)action; (void)attr; @@ -1634,17 +1817,16 @@ flow_dv_validate_action_pop_vlan(struct rte_eth_dev *dev, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, "pop vlan action is not supported"); - /* - * Check for inconsistencies: - * fail strip_vlan in a flow that matches packets without VLAN tags. - * fail strip_vlan in a flow that matches packets without explicitly a - * matching on VLAN tag ? - */ - if (action_flags & MLX5_FLOW_ACTION_OF_POP_VLAN) + if (attr->egress) return rte_flow_error_set(error, ENOTSUP, - RTE_FLOW_ERROR_TYPE_UNSPECIFIED, + RTE_FLOW_ERROR_TYPE_ATTR_EGRESS, NULL, - "no support for multiple vlan pop " + "pop vlan action not supported for " + "egress"); + if (action_flags & MLX5_FLOW_VLAN_ACTIONS) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ACTION, action, + "no support for multiple VLAN " "actions"); if (!(item_flags & MLX5_FLOW_LAYER_OUTER_VLAN)) return rte_flow_error_set(error, ENOTSUP, @@ -1657,20 +1839,21 @@ flow_dv_validate_action_pop_vlan(struct rte_eth_dev *dev, RTE_FLOW_ERROR_TYPE_ACTION, action, "wrong action order, port_id should " "be after pop VLAN action"); + if (!attr->transfer && priv->representor) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, + "pop vlan action for VF representor " + "not supported on NIC table"); return 0; } /** * Get VLAN default info from vlan match info. * - * @param[in] dev - * Pointer to the rte_eth_dev structure. - * @param[in] item + * @param[in] items * the list of item specifications. * @param[out] vlan * pointer VLAN info to fill to. - * @param[out] error - * Pointer to error structure. * * @return * 0 on success, a negative errno value otherwise and rte_errno is set. @@ -1687,10 +1870,14 @@ flow_dev_get_vlan_info_from_items(const struct rte_flow_item *items, if (items == NULL) return; - for (; items->type != RTE_FLOW_ITEM_TYPE_END && - items->type != RTE_FLOW_ITEM_TYPE_VLAN; items++) - ; - if (items->type == RTE_FLOW_ITEM_TYPE_VLAN) { + for (; items->type != RTE_FLOW_ITEM_TYPE_END; items++) { + int type = items->type; + + if (type == RTE_FLOW_ITEM_TYPE_VLAN || + type == MLX5_RTE_FLOW_ITEM_TYPE_VLAN) + break; + } + if (items->type != RTE_FLOW_ITEM_TYPE_END) { const struct rte_flow_item_vlan *vlan_m = items->mask; const struct rte_flow_item_vlan *vlan_v = items->spec; @@ -1699,7 +1886,7 @@ flow_dev_get_vlan_info_from_items(const struct rte_flow_item *items, /* Only full match values are accepted */ if ((vlan_m->tci & MLX5DV_FLOW_VLAN_PCP_MASK_BE) == MLX5DV_FLOW_VLAN_PCP_MASK_BE) { - vlan->vlan_tci &= MLX5DV_FLOW_VLAN_PCP_MASK; + vlan->vlan_tci &= ~MLX5DV_FLOW_VLAN_PCP_MASK; vlan->vlan_tci |= rte_be_to_cpu_16(vlan_v->tci & MLX5DV_FLOW_VLAN_PCP_MASK_BE); @@ -1720,10 +1907,14 @@ flow_dev_get_vlan_info_from_items(const struct rte_flow_item *items, /** * Validate the push VLAN action. * + * @param[in] dev + * Pointer to the rte_eth_dev structure. * @param[in] action_flags * Holds the actions detected until now. + * @param[in] item_flags + * The items found in this flow rule. * @param[in] action - * Pointer to the encap action. + * Pointer to the action structure. * @param[in] attr * Pointer to flow attributes * @param[out] error @@ -1733,14 +1924,22 @@ flow_dev_get_vlan_info_from_items(const struct rte_flow_item *items, * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -flow_dv_validate_action_push_vlan(uint64_t action_flags, - uint64_t item_flags, +flow_dv_validate_action_push_vlan(struct rte_eth_dev *dev, + uint64_t action_flags, + const struct rte_flow_item_vlan *vlan_m, const struct rte_flow_action *action, const struct rte_flow_attr *attr, struct rte_flow_error *error) { const struct rte_flow_action_of_push_vlan *push_vlan = action->conf; + const struct mlx5_priv *priv = dev->data->dev_private; + if (!attr->transfer && attr->ingress) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ATTR_INGRESS, + NULL, + "push VLAN action not supported for " + "ingress"); if (push_vlan->ethertype != RTE_BE16(RTE_ETHER_TYPE_VLAN) && push_vlan->ethertype != RTE_BE16(RTE_ETHER_TYPE_QINQ)) return rte_flow_error_set(error, EINVAL, @@ -1751,19 +1950,42 @@ flow_dv_validate_action_push_vlan(uint64_t action_flags, RTE_FLOW_ERROR_TYPE_ACTION, action, "no support for multiple VLAN " "actions"); - if (!mlx5_flow_find_action - (action + 1, RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_VID) && - !(item_flags & MLX5_FLOW_LAYER_OUTER_VLAN)) - return rte_flow_error_set(error, ENOTSUP, - RTE_FLOW_ERROR_TYPE_ACTION, action, - "push VLAN needs to match on VLAN in order to " - "get VLAN VID information because there is " - "no followed set VLAN VID action"); if (action_flags & MLX5_FLOW_ACTION_PORT_ID) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, action, "wrong action order, port_id should " "be after push VLAN"); + if (!attr->transfer && priv->representor) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, + "push vlan action for VF representor " + "not supported on NIC table"); + if (vlan_m && + (vlan_m->tci & MLX5DV_FLOW_VLAN_PCP_MASK_BE) && + (vlan_m->tci & MLX5DV_FLOW_VLAN_PCP_MASK_BE) != + MLX5DV_FLOW_VLAN_PCP_MASK_BE && + !(action_flags & MLX5_FLOW_ACTION_OF_SET_VLAN_PCP) && + !(mlx5_flow_find_action + (action + 1, RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_PCP))) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, action, + "not full match mask on VLAN PCP and " + "there is no of_set_vlan_pcp action, " + "push VLAN action cannot figure out " + "PCP value"); + if (vlan_m && + (vlan_m->tci & MLX5DV_FLOW_VLAN_VID_MASK_BE) && + (vlan_m->tci & MLX5DV_FLOW_VLAN_VID_MASK_BE) != + MLX5DV_FLOW_VLAN_VID_MASK_BE && + !(action_flags & MLX5_FLOW_ACTION_OF_SET_VLAN_VID) && + !(mlx5_flow_find_action + (action + 1, RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_VID))) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, action, + "not full match mask on VLAN VID and " + "there is no of_set_vlan_vid action, " + "push VLAN action cannot figure out " + "VID value"); (void)attr; return 0; } @@ -1775,8 +1997,6 @@ flow_dv_validate_action_push_vlan(uint64_t action_flags, * Holds the actions detected until now. * @param[in] actions * Pointer to the list of actions remaining in the flow rule. - * @param[in] attr - * Pointer to flow attributes * @param[out] error * Pointer to error structure. * @@ -1818,10 +2038,10 @@ flow_dv_validate_action_set_vlan_pcp(uint64_t action_flags, * * @param[in] item_flags * Holds the items detected in this rule. + * @param[in] action_flags + * Holds the actions detected until now. * @param[in] actions * Pointer to the list of actions remaining in the flow rule. - * @param[in] attr - * Pointer to flow attributes * @param[out] error * Pointer to error structure. * @@ -1837,37 +2057,21 @@ flow_dv_validate_action_set_vlan_vid(uint64_t item_flags, const struct rte_flow_action *action = actions; const struct rte_flow_action_of_set_vlan_vid *conf = action->conf; - if (conf->vlan_vid > RTE_BE16(0xFFE)) + if (rte_be_to_cpu_16(conf->vlan_vid) > 0xFFE) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, action, "VLAN VID value is too big"); - /* there is an of_push_vlan action before us */ - if (action_flags & MLX5_FLOW_ACTION_OF_PUSH_VLAN) { - if (mlx5_flow_find_action(actions + 1, - RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_VID)) - return rte_flow_error_set(error, ENOTSUP, - RTE_FLOW_ERROR_TYPE_ACTION, action, - "Multiple VLAN VID modifications are " - "not supported"); - else - return 0; - } - - /* - * Action is on an existing VLAN header: - * Need to verify this is a single modify CID action. - * Rule mast include a match on outer VLAN. - */ + if (!(action_flags & MLX5_FLOW_ACTION_OF_PUSH_VLAN) && + !(item_flags & MLX5_FLOW_LAYER_OUTER_VLAN)) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ACTION, action, + "set VLAN VID action must follow push" + " VLAN action or match on VLAN item"); if (action_flags & MLX5_FLOW_ACTION_OF_SET_VLAN_VID) return rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ACTION, action, "Multiple VLAN VID modifications are " "not supported"); - if (!(item_flags & MLX5_FLOW_LAYER_OUTER_VLAN)) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, action, - "match on VLAN is required in order " - "to set VLAN VID"); if (action_flags & MLX5_FLOW_ACTION_PORT_ID) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, action, @@ -1919,11 +2123,7 @@ flow_dv_validate_action_flag(struct rte_eth_dev *dev, ret = mlx5_flow_get_reg_id(dev, MLX5_FLOW_MARK, 0, error); if (ret < 0) return ret; - assert(ret > 0); - if (action_flags & MLX5_FLOW_ACTION_DROP) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, NULL, - "can't drop and flag in same flow"); + MLX5_ASSERT(ret > 0); if (action_flags & MLX5_FLOW_ACTION_MARK) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, @@ -1983,7 +2183,7 @@ flow_dv_validate_action_mark(struct rte_eth_dev *dev, ret = mlx5_flow_get_reg_id(dev, MLX5_FLOW_MARK, 0, error); if (ret < 0) return ret; - assert(ret > 0); + MLX5_ASSERT(ret > 0); if (!mark) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, action, @@ -1993,10 +2193,6 @@ flow_dv_validate_action_mark(struct rte_eth_dev *dev, RTE_FLOW_ERROR_TYPE_ACTION_CONF, &mark->id, "mark id exceeds the limit"); - if (action_flags & MLX5_FLOW_ACTION_DROP) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, NULL, - "can't drop and mark in same flow"); if (action_flags & MLX5_FLOW_ACTION_FLAG) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, @@ -2015,7 +2211,7 @@ flow_dv_validate_action_mark(struct rte_eth_dev *dev, * @param[in] dev * Pointer to the rte_eth_dev structure. * @param[in] action - * Pointer to the encap action. + * Pointer to the action structure. * @param[in] action_flags * Holds the actions detected until now. * @param[in] attr @@ -2035,7 +2231,7 @@ flow_dv_validate_action_set_meta(struct rte_eth_dev *dev, { const struct rte_flow_action_set_meta *conf; uint32_t nic_mask = UINT32_MAX; - enum modify_reg reg; + int reg; if (!mlx5_flow_ext_mreg_supported(dev)) return rte_flow_error_set(error, ENOTSUP, @@ -2063,10 +2259,6 @@ flow_dv_validate_action_set_meta(struct rte_eth_dev *dev, return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, action, "meta data must be within reg C0"); - if (!(conf->data & conf->mask)) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, action, - "zero value has no effect"); return 0; } @@ -2076,7 +2268,7 @@ flow_dv_validate_action_set_meta(struct rte_eth_dev *dev, * @param[in] dev * Pointer to the rte_eth_dev structure. * @param[in] action - * Pointer to the encap action. + * Pointer to the action structure. * @param[in] action_flags * Holds the actions detected until now. * @param[in] attr @@ -2130,7 +2322,7 @@ flow_dv_validate_action_set_tag(struct rte_eth_dev *dev, * Validate count action. * * @param[in] dev - * device otr. + * Pointer to rte_eth_dev structure. * @param[out] error * Pointer to error structure. * @@ -2159,12 +2351,14 @@ notsup_err: /** * Validate the L2 encap action. * + * @param[in] dev + * Pointer to the rte_eth_dev structure. * @param[in] action_flags * Holds the actions detected until now. * @param[in] action - * Pointer to the encap action. + * Pointer to the action structure. * @param[in] attr - * Pointer to flow attributes + * Pointer to flow attributes. * @param[out] error * Pointer to error structure. * @@ -2172,36 +2366,36 @@ notsup_err: * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -flow_dv_validate_action_l2_encap(uint64_t action_flags, +flow_dv_validate_action_l2_encap(struct rte_eth_dev *dev, + uint64_t action_flags, const struct rte_flow_action *action, const struct rte_flow_attr *attr, struct rte_flow_error *error) { + const struct mlx5_priv *priv = dev->data->dev_private; + if (!(action->conf)) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, action, "configuration cannot be null"); - if (action_flags & MLX5_FLOW_ACTION_DROP) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, NULL, - "can't drop and encap in same flow"); - if (action_flags & (MLX5_FLOW_ENCAP_ACTIONS | MLX5_FLOW_DECAP_ACTIONS)) + if (action_flags & MLX5_FLOW_ACTION_ENCAP) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, - "can only have a single encap or" - " decap action in a flow"); - if (!attr->transfer && attr->ingress) + "can only have a single encap action " + "in a flow"); + if (!attr->transfer && priv->representor) return rte_flow_error_set(error, ENOTSUP, - RTE_FLOW_ERROR_TYPE_ATTR_INGRESS, - NULL, - "encap action not supported for " - "ingress"); + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, + "encap action for VF representor " + "not supported on NIC table"); return 0; } /** - * Validate the L2 decap action. + * Validate a decap action. * + * @param[in] dev + * Pointer to the rte_eth_dev structure. * @param[in] action_flags * Holds the actions detected until now. * @param[in] attr @@ -2213,19 +2407,20 @@ flow_dv_validate_action_l2_encap(uint64_t action_flags, * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -flow_dv_validate_action_l2_decap(uint64_t action_flags, - const struct rte_flow_attr *attr, - struct rte_flow_error *error) +flow_dv_validate_action_decap(struct rte_eth_dev *dev, + uint64_t action_flags, + const struct rte_flow_attr *attr, + struct rte_flow_error *error) { - if (action_flags & MLX5_FLOW_ACTION_DROP) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, NULL, - "can't drop and decap in same flow"); - if (action_flags & (MLX5_FLOW_ENCAP_ACTIONS | MLX5_FLOW_DECAP_ACTIONS)) - return rte_flow_error_set(error, EINVAL, + const struct mlx5_priv *priv = dev->data->dev_private; + + if (action_flags & MLX5_FLOW_XCAP_ACTIONS) + return rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ACTION, NULL, - "can only have a single encap or" - " decap action in a flow"); + action_flags & + MLX5_FLOW_ACTION_DECAP ? "can only " + "have a single decap action" : "decap " + "after encap is not supported"); if (action_flags & MLX5_FLOW_MODIFY_HDR_ACTIONS) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, @@ -2237,18 +2432,31 @@ flow_dv_validate_action_l2_decap(uint64_t action_flags, NULL, "decap action not supported for " "egress"); + if (!attr->transfer && priv->representor) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, + "decap action for VF representor " + "not supported on NIC table"); return 0; } +const struct rte_flow_action_raw_decap empty_decap = {.data = NULL, .size = 0,}; + /** - * Validate the raw encap action. + * Validate the raw encap and decap actions. * - * @param[in] action_flags - * Holds the actions detected until now. - * @param[in] action + * @param[in] dev + * Pointer to the rte_eth_dev structure. + * @param[in] decap + * Pointer to the decap action. + * @param[in] encap * Pointer to the encap action. * @param[in] attr * Pointer to flow attributes + * @param[in/out] action_flags + * Holds the actions detected until now. + * @param[out] actions_n + * pointer to the number of actions counter. * @param[out] error * Pointer to error structure. * @@ -2256,110 +2464,90 @@ flow_dv_validate_action_l2_decap(uint64_t action_flags, * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -flow_dv_validate_action_raw_encap(uint64_t action_flags, - const struct rte_flow_action *action, - const struct rte_flow_attr *attr, - struct rte_flow_error *error) -{ - const struct rte_flow_action_raw_encap *raw_encap = - (const struct rte_flow_action_raw_encap *)action->conf; - if (!(action->conf)) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, action, - "configuration cannot be null"); - if (action_flags & MLX5_FLOW_ACTION_DROP) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, NULL, - "can't drop and encap in same flow"); - if (action_flags & MLX5_FLOW_ENCAP_ACTIONS) +flow_dv_validate_action_raw_encap_decap + (struct rte_eth_dev *dev, + const struct rte_flow_action_raw_decap *decap, + const struct rte_flow_action_raw_encap *encap, + const struct rte_flow_attr *attr, uint64_t *action_flags, + int *actions_n, struct rte_flow_error *error) +{ + const struct mlx5_priv *priv = dev->data->dev_private; + int ret; + + if (encap && (!encap->size || !encap->data)) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, - "can only have a single encap" - " action in a flow"); - /* encap without preceding decap is not supported for ingress */ - if (!attr->transfer && attr->ingress && - !(action_flags & MLX5_FLOW_ACTION_RAW_DECAP)) - return rte_flow_error_set(error, ENOTSUP, - RTE_FLOW_ERROR_TYPE_ATTR_INGRESS, - NULL, - "encap action not supported for " - "ingress"); - if (!raw_encap->size || !raw_encap->data) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, action, "raw encap data cannot be empty"); + if (decap && encap) { + if (decap->size <= MLX5_ENCAPSULATION_DECISION_SIZE && + encap->size > MLX5_ENCAPSULATION_DECISION_SIZE) + /* L3 encap. */ + decap = NULL; + else if (encap->size <= + MLX5_ENCAPSULATION_DECISION_SIZE && + decap->size > + MLX5_ENCAPSULATION_DECISION_SIZE) + /* L3 decap. */ + encap = NULL; + else if (encap->size > + MLX5_ENCAPSULATION_DECISION_SIZE && + decap->size > + MLX5_ENCAPSULATION_DECISION_SIZE) + /* 2 L2 actions: encap and decap. */ + ; + else + return rte_flow_error_set(error, + ENOTSUP, + RTE_FLOW_ERROR_TYPE_ACTION, + NULL, "unsupported too small " + "raw decap and too small raw " + "encap combination"); + } + if (decap) { + ret = flow_dv_validate_action_decap(dev, *action_flags, attr, + error); + if (ret < 0) + return ret; + *action_flags |= MLX5_FLOW_ACTION_DECAP; + ++(*actions_n); + } + if (encap) { + if (encap->size <= MLX5_ENCAPSULATION_DECISION_SIZE) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ACTION, + NULL, + "small raw encap size"); + if (*action_flags & MLX5_FLOW_ACTION_ENCAP) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, + NULL, + "more than one encap action"); + if (!attr->transfer && priv->representor) + return rte_flow_error_set + (error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, + "encap action for VF representor " + "not supported on NIC table"); + *action_flags |= MLX5_FLOW_ACTION_ENCAP; + ++(*actions_n); + } return 0; } /** - * Validate the raw decap action. + * Find existing encap/decap resource or create and register a new one. * - * @param[in] action_flags - * Holds the actions detected until now. - * @param[in] action - * Pointer to the encap action. - * @param[in] attr - * Pointer to flow attributes + * @param[in, out] dev + * Pointer to rte_eth_dev structure. + * @param[in, out] resource + * Pointer to encap/decap resource. + * @parm[in, out] dev_flow + * Pointer to the dev_flow. * @param[out] error - * Pointer to error structure. + * pointer to error structure. * * @return - * 0 on success, a negative errno value otherwise and rte_errno is set. - */ -static int -flow_dv_validate_action_raw_decap(uint64_t action_flags, - const struct rte_flow_action *action, - const struct rte_flow_attr *attr, - struct rte_flow_error *error) -{ - const struct rte_flow_action_raw_decap *decap = action->conf; - - if (action_flags & MLX5_FLOW_ACTION_DROP) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, NULL, - "can't drop and decap in same flow"); - if (action_flags & MLX5_FLOW_ENCAP_ACTIONS) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, NULL, - "can't have encap action before" - " decap action"); - if (action_flags & MLX5_FLOW_DECAP_ACTIONS) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, NULL, - "can only have a single decap" - " action in a flow"); - /* decap action is valid on egress only if it is followed by encap */ - if (attr->egress && decap && - decap->size > MLX5_ENCAPSULATION_DECISION_SIZE) { - return rte_flow_error_set(error, ENOTSUP, - RTE_FLOW_ERROR_TYPE_ATTR_EGRESS, - NULL, "decap action not supported" - " for egress"); - } else if (decap && decap->size > MLX5_ENCAPSULATION_DECISION_SIZE && - (action_flags & MLX5_FLOW_MODIFY_HDR_ACTIONS)) { - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION, - NULL, - "can't have decap action " - "after modify action"); - } - return 0; -} - -/** - * Find existing encap/decap resource or create and register a new one. - * - * @param[in, out] dev - * Pointer to rte_eth_dev structure. - * @param[in, out] resource - * Pointer to encap/decap resource. - * @parm[in, out] dev_flow - * Pointer to the dev_flow. - * @param[out] error - * pointer to error structure. - * - * @return - * 0 on success otherwise -errno and errno is set. + * 0 on success otherwise -errno and errno is set. */ static int flow_dv_encap_decap_resource_register @@ -2372,8 +2560,9 @@ flow_dv_encap_decap_resource_register struct mlx5_ibv_shared *sh = priv->sh; struct mlx5_flow_dv_encap_decap_resource *cache_resource; struct mlx5dv_dr_domain *domain; + uint32_t idx = 0; - resource->flags = dev_flow->group ? 0 : 1; + resource->flags = dev_flow->dv.group ? 0 : 1; if (resource->ft_type == MLX5DV_FLOW_TABLE_TYPE_FDB) domain = sh->fdb_domain; else if (resource->ft_type == MLX5DV_FLOW_TABLE_TYPE_NIC_RX) @@ -2381,7 +2570,8 @@ flow_dv_encap_decap_resource_register else domain = sh->tx_domain; /* Lookup a matching resource from cache. */ - LIST_FOREACH(cache_resource, &sh->encaps_decaps, next) { + ILIST_FOREACH(sh->ipool[MLX5_IPOOL_DECAP_ENCAP], sh->encaps_decaps, idx, + cache_resource, next) { if (resource->reformat_type == cache_resource->reformat_type && resource->ft_type == cache_resource->ft_type && resource->flags == cache_resource->flags && @@ -2393,12 +2583,14 @@ flow_dv_encap_decap_resource_register (void *)cache_resource, rte_atomic32_read(&cache_resource->refcnt)); rte_atomic32_inc(&cache_resource->refcnt); + dev_flow->handle->dvh.rix_encap_decap = idx; dev_flow->dv.encap_decap = cache_resource; return 0; } } /* Register new encap/decap resource. */ - cache_resource = rte_calloc(__func__, 1, sizeof(*cache_resource), 0); + cache_resource = mlx5_ipool_zmalloc(sh->ipool[MLX5_IPOOL_DECAP_ENCAP], + &dev_flow->handle->dvh.rix_encap_decap); if (!cache_resource) return rte_flow_error_set(error, ENOMEM, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, @@ -2418,7 +2610,9 @@ flow_dv_encap_decap_resource_register } rte_atomic32_init(&cache_resource->refcnt); rte_atomic32_inc(&cache_resource->refcnt); - LIST_INSERT_HEAD(&sh->encaps_decaps, cache_resource, next); + ILIST_INSERT(sh->ipool[MLX5_IPOOL_DECAP_ENCAP], &sh->encaps_decaps, + dev_flow->handle->dvh.rix_encap_decap, cache_resource, + next); dev_flow->dv.encap_decap = cache_resource; DRV_LOG(DEBUG, "new encap/decap resource %p: refcnt %d++", (void *)cache_resource, @@ -2452,7 +2646,7 @@ flow_dv_jump_tbl_resource_register container_of(tbl, struct mlx5_flow_tbl_data_entry, tbl); int cnt; - assert(tbl); + MLX5_ASSERT(tbl); cnt = rte_atomic32_read(&tbl_data->jump.refcnt); if (!cnt) { tbl_data->jump.action = @@ -2465,11 +2659,14 @@ flow_dv_jump_tbl_resource_register DRV_LOG(DEBUG, "new jump table resource %p: refcnt %d++", (void *)&tbl_data->jump, cnt); } else { - assert(tbl_data->jump.action); + /* old jump should not make the table ref++. */ + flow_dv_tbl_resource_release(dev, &tbl_data->tbl); + MLX5_ASSERT(tbl_data->jump.action); DRV_LOG(DEBUG, "existed jump table resource %p: refcnt %d++", (void *)&tbl_data->jump, cnt); } rte_atomic32_inc(&tbl_data->jump.refcnt); + dev_flow->handle->rix_jump = tbl_data->idx; dev_flow->dv.jump = &tbl_data->jump; return 0; } @@ -2499,21 +2696,25 @@ flow_dv_port_id_action_resource_register struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_ibv_shared *sh = priv->sh; struct mlx5_flow_dv_port_id_action_resource *cache_resource; + uint32_t idx = 0; /* Lookup a matching resource from cache. */ - LIST_FOREACH(cache_resource, &sh->port_id_action_list, next) { + ILIST_FOREACH(sh->ipool[MLX5_IPOOL_PORT_ID], sh->port_id_action_list, + idx, cache_resource, next) { if (resource->port_id == cache_resource->port_id) { DRV_LOG(DEBUG, "port id action resource resource %p: " "refcnt %d++", (void *)cache_resource, rte_atomic32_read(&cache_resource->refcnt)); rte_atomic32_inc(&cache_resource->refcnt); + dev_flow->handle->rix_port_id_action = idx; dev_flow->dv.port_id_action = cache_resource; return 0; } } /* Register new port id action resource. */ - cache_resource = rte_calloc(__func__, 1, sizeof(*cache_resource), 0); + cache_resource = mlx5_ipool_zmalloc(sh->ipool[MLX5_IPOOL_PORT_ID], + &dev_flow->handle->rix_port_id_action); if (!cache_resource) return rte_flow_error_set(error, ENOMEM, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, @@ -2535,7 +2736,9 @@ flow_dv_port_id_action_resource_register } rte_atomic32_init(&cache_resource->refcnt); rte_atomic32_inc(&cache_resource->refcnt); - LIST_INSERT_HEAD(&sh->port_id_action_list, cache_resource, next); + ILIST_INSERT(sh->ipool[MLX5_IPOOL_PORT_ID], &sh->port_id_action_list, + dev_flow->handle->rix_port_id_action, cache_resource, + next); dev_flow->dv.port_id_action = cache_resource; DRV_LOG(DEBUG, "new port id action resource %p: refcnt %d++", (void *)cache_resource, @@ -2569,9 +2772,11 @@ flow_dv_push_vlan_action_resource_register struct mlx5_ibv_shared *sh = priv->sh; struct mlx5_flow_dv_push_vlan_action_resource *cache_resource; struct mlx5dv_dr_domain *domain; + uint32_t idx = 0; /* Lookup a matching resource from cache. */ - LIST_FOREACH(cache_resource, &sh->push_vlan_action_list, next) { + ILIST_FOREACH(sh->ipool[MLX5_IPOOL_PUSH_VLAN], + sh->push_vlan_action_list, idx, cache_resource, next) { if (resource->vlan_tag == cache_resource->vlan_tag && resource->ft_type == cache_resource->ft_type) { DRV_LOG(DEBUG, "push-VLAN action resource resource %p: " @@ -2579,12 +2784,14 @@ flow_dv_push_vlan_action_resource_register (void *)cache_resource, rte_atomic32_read(&cache_resource->refcnt)); rte_atomic32_inc(&cache_resource->refcnt); + dev_flow->handle->dvh.rix_push_vlan = idx; dev_flow->dv.push_vlan_res = cache_resource; return 0; } } /* Register new push_vlan action resource. */ - cache_resource = rte_calloc(__func__, 1, sizeof(*cache_resource), 0); + cache_resource = mlx5_ipool_zmalloc(sh->ipool[MLX5_IPOOL_PUSH_VLAN], + &dev_flow->handle->dvh.rix_push_vlan); if (!cache_resource) return rte_flow_error_set(error, ENOMEM, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, @@ -2607,7 +2814,10 @@ flow_dv_push_vlan_action_resource_register } rte_atomic32_init(&cache_resource->refcnt); rte_atomic32_inc(&cache_resource->refcnt); - LIST_INSERT_HEAD(&sh->push_vlan_action_list, cache_resource, next); + ILIST_INSERT(sh->ipool[MLX5_IPOOL_PUSH_VLAN], + &sh->push_vlan_action_list, + dev_flow->handle->dvh.rix_push_vlan, + cache_resource, next); dev_flow->dv.push_vlan_res = cache_resource; DRV_LOG(DEBUG, "new push vlan action resource %p: refcnt %d++", (void *)cache_resource, @@ -2930,8 +3140,6 @@ flow_dv_create_action_l2_encap(struct rte_eth_dev *dev, (const struct rte_flow_action_raw_encap *)action->conf; res.size = raw_encap_data->size; memcpy(res.buf, raw_encap_data->data, res.size); - if (flow_dv_zero_encap_udp_csum(res.buf, error)) - return -rte_errno; } else { if (action->type == RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP) encap_data = @@ -2945,6 +3153,8 @@ flow_dv_create_action_l2_encap(struct rte_eth_dev *dev, &res.size, error)) return -rte_errno; } + if (flow_dv_zero_encap_udp_csum(res.buf, error)) + return -rte_errno; if (flow_dv_encap_decap_resource_register(dev, &res, dev_flow, error)) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, @@ -3015,6 +3225,7 @@ flow_dv_create_action_raw_encap(struct rte_eth_dev *dev, const struct rte_flow_action_raw_encap *encap_data; struct mlx5_flow_dv_encap_decap_resource res; + memset(&res, 0, sizeof(res)); encap_data = (const struct rte_flow_action_raw_encap *)action->conf; res.size = encap_data->size; memcpy(res.buf, encap_data->data, res.size); @@ -3038,12 +3249,12 @@ flow_dv_create_action_raw_encap(struct rte_eth_dev *dev, * * @param[in] dev * Pointer to rte_eth_dev structure. - * @param[in] vlan_tag - * the vlan tag to push to the Ethernet header. - * @param[in, out] dev_flow - * Pointer to the mlx5_flow. * @param[in] attr * Pointer to the flow attributes. + * @param[in] vlan + * Pointer to the vlan to push to the Ethernet header. + * @param[in, out] dev_flow + * Pointer to the mlx5_flow. * @param[out] error * Pointer to the error structure. * @@ -3059,6 +3270,7 @@ flow_dv_create_action_push_vlan(struct rte_eth_dev *dev, { struct mlx5_flow_dv_push_vlan_action_resource res; + memset(&res, 0, sizeof(res)); res.vlan_tag = rte_cpu_to_be_32(((uint32_t)vlan->eth_proto) << 16 | vlan->vlan_tci); @@ -3093,7 +3305,7 @@ flow_dv_validate_action_modify_hdr(const uint64_t action_flags, return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION_CONF, NULL, "action configuration not set"); - if (action_flags & MLX5_FLOW_ENCAP_ACTIONS) + if (action_flags & MLX5_FLOW_ACTION_ENCAP) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, "can't have encap action before" @@ -3157,10 +3369,14 @@ flow_dv_validate_action_modify_ipv4(const uint64_t action_flags, struct rte_flow_error *error) { int ret = 0; + uint64_t layer; ret = flow_dv_validate_action_modify_hdr(action_flags, action, error); if (!ret) { - if (!(item_flags & MLX5_FLOW_LAYER_L3_IPV4)) + layer = (action_flags & MLX5_FLOW_ACTION_DECAP) ? + MLX5_FLOW_LAYER_INNER_L3_IPV4 : + MLX5_FLOW_LAYER_OUTER_L3_IPV4; + if (!(item_flags & layer)) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, @@ -3191,10 +3407,14 @@ flow_dv_validate_action_modify_ipv6(const uint64_t action_flags, struct rte_flow_error *error) { int ret = 0; + uint64_t layer; ret = flow_dv_validate_action_modify_hdr(action_flags, action, error); if (!ret) { - if (!(item_flags & MLX5_FLOW_LAYER_L3_IPV6)) + layer = (action_flags & MLX5_FLOW_ACTION_DECAP) ? + MLX5_FLOW_LAYER_INNER_L3_IPV6 : + MLX5_FLOW_LAYER_OUTER_L3_IPV6; + if (!(item_flags & layer)) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, @@ -3225,10 +3445,14 @@ flow_dv_validate_action_modify_tp(const uint64_t action_flags, struct rte_flow_error *error) { int ret = 0; + uint64_t layer; ret = flow_dv_validate_action_modify_hdr(action_flags, action, error); if (!ret) { - if (!(item_flags & MLX5_FLOW_LAYER_L4)) + layer = (action_flags & MLX5_FLOW_ACTION_DECAP) ? + MLX5_FLOW_LAYER_INNER_L4 : + MLX5_FLOW_LAYER_OUTER_L4; + if (!(item_flags & layer)) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, "no transport layer " @@ -3260,10 +3484,14 @@ flow_dv_validate_action_modify_tcp_seq(const uint64_t action_flags, struct rte_flow_error *error) { int ret = 0; + uint64_t layer; ret = flow_dv_validate_action_modify_hdr(action_flags, action, error); if (!ret) { - if (!(item_flags & MLX5_FLOW_LAYER_OUTER_L4_TCP)) + layer = (action_flags & MLX5_FLOW_ACTION_DECAP) ? + MLX5_FLOW_LAYER_INNER_L4_TCP : + MLX5_FLOW_LAYER_OUTER_L4_TCP; + if (!(item_flags & layer)) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, "no TCP item in" @@ -3305,10 +3533,14 @@ flow_dv_validate_action_modify_tcp_ack(const uint64_t action_flags, struct rte_flow_error *error) { int ret = 0; + uint64_t layer; ret = flow_dv_validate_action_modify_hdr(action_flags, action, error); if (!ret) { - if (!(item_flags & MLX5_FLOW_LAYER_OUTER_L4_TCP)) + layer = (action_flags & MLX5_FLOW_ACTION_DECAP) ? + MLX5_FLOW_LAYER_INNER_L4_TCP : + MLX5_FLOW_LAYER_OUTER_L4_TCP; + if (!(item_flags & layer)) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, "no TCP item in" @@ -3349,10 +3581,14 @@ flow_dv_validate_action_modify_ttl(const uint64_t action_flags, struct rte_flow_error *error) { int ret = 0; + uint64_t layer; ret = flow_dv_validate_action_modify_hdr(action_flags, action, error); if (!ret) { - if (!(item_flags & MLX5_FLOW_LAYER_L3)) + layer = (action_flags & MLX5_FLOW_ACTION_DECAP) ? + MLX5_FLOW_LAYER_INNER_L3 : + MLX5_FLOW_LAYER_OUTER_L3; + if (!(item_flags & layer)) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, @@ -3404,7 +3640,7 @@ flow_dv_validate_action_jump(const struct rte_flow_action *action, target_group = ((const struct rte_flow_action_jump *)action->conf)->group; ret = mlx5_flow_group_to_table(attributes, external, target_group, - &table, error); + true, &table, error); if (ret) return ret; if (attributes->group == target_group) @@ -3496,21 +3732,18 @@ flow_dv_validate_action_port_id(struct rte_eth_dev *dev, * @return * Max number of modify header actions device can support. */ -static unsigned int -flow_dv_modify_hdr_action_max(struct rte_eth_dev *dev, uint64_t flags) +static inline unsigned int +flow_dv_modify_hdr_action_max(struct rte_eth_dev *dev __rte_unused, + uint64_t flags) { /* - * There's no way to directly query the max cap. Although it has to be - * acquried by iterative trial, it is a safe assumption that more - * actions are supported by FW if extensive metadata register is - * supported. (Only in the root table) + * There's no way to directly query the max capacity from FW. + * The maximal value on root table should be assumed to be supported. */ if (!(flags & MLX5DV_DR_ACTION_FLAGS_ROOT_LEVEL)) return MLX5_MAX_MODIFY_NUM; else - return mlx5_flow_ext_mreg_supported(dev) ? - MLX5_ROOT_TBL_MODIFY_NUM : - MLX5_ROOT_TBL_MODIFY_NUM_NO_MREG; + return MLX5_ROOT_TBL_MODIFY_NUM; } /** @@ -3564,9 +3797,9 @@ mlx5_flow_validate_action_meter(struct rte_eth_dev *dev, return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, "Meter not found"); - if (fm->ref_cnt && (!(fm->attr.transfer == attr->transfer || - (!fm->attr.ingress && !attr->ingress && attr->egress) || - (!fm->attr.egress && !attr->egress && attr->ingress)))) + if (fm->ref_cnt && (!(fm->transfer == attr->transfer || + (!fm->ingress && !attr->ingress && attr->egress) || + (!fm->egress && !attr->egress && attr->ingress)))) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, NULL, "Flow attributes are either invalid " @@ -3575,6 +3808,50 @@ mlx5_flow_validate_action_meter(struct rte_eth_dev *dev, return 0; } +/** + * Validate the age action. + * + * @param[in] action_flags + * Holds the actions detected until now. + * @param[in] action + * Pointer to the age action. + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[out] error + * Pointer to error structure. + * + * @return + * 0 on success, a negative errno value otherwise and rte_errno is set. + */ +static int +flow_dv_validate_action_age(uint64_t action_flags, + const struct rte_flow_action *action, + struct rte_eth_dev *dev, + struct rte_flow_error *error) +{ + struct mlx5_priv *priv = dev->data->dev_private; + const struct rte_flow_action_age *age = action->conf; + + if (!priv->config.devx || priv->counter_fallback) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, + NULL, + "age action not supported"); + if (!(action->conf)) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, action, + "configuration cannot be null"); + if (age->timeout >= UINT16_MAX / 2 / 10) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ACTION, action, + "Max age time: 3275 seconds"); + if (action_flags & MLX5_FLOW_ACTION_AGE) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, NULL, + "Duplicate age ctions set"); + return 0; +} + /** * Validate the modify-header IPv4 DSCP actions. * @@ -3671,8 +3948,8 @@ flow_dv_modify_hdr_resource_register struct mlx5dv_dr_domain *ns; uint32_t actions_len; - resource->flags = - dev_flow->group ? 0 : MLX5DV_DR_ACTION_FLAGS_ROOT_LEVEL; + resource->flags = dev_flow->dv.group ? 0 : + MLX5DV_DR_ACTION_FLAGS_ROOT_LEVEL; if (resource->actions_num > flow_dv_modify_hdr_action_max(dev, resource->flags)) return rte_flow_error_set(error, EOVERFLOW, @@ -3697,7 +3974,7 @@ flow_dv_modify_hdr_resource_register (void *)cache_resource, rte_atomic32_read(&cache_resource->refcnt)); rte_atomic32_inc(&cache_resource->refcnt); - dev_flow->dv.modify_hdr = cache_resource; + dev_flow->handle->dvh.modify_hdr = cache_resource; return 0; } } @@ -3724,139 +4001,50 @@ flow_dv_modify_hdr_resource_register rte_atomic32_init(&cache_resource->refcnt); rte_atomic32_inc(&cache_resource->refcnt); LIST_INSERT_HEAD(&sh->modify_cmds, cache_resource, next); - dev_flow->dv.modify_hdr = cache_resource; + dev_flow->handle->dvh.modify_hdr = cache_resource; DRV_LOG(DEBUG, "new modify-header resource %p: refcnt %d++", (void *)cache_resource, rte_atomic32_read(&cache_resource->refcnt)); return 0; } -#define MLX5_CNT_CONTAINER_RESIZE 64 - /** - * Get or create a flow counter. + * Get DV flow counter by index. * * @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. + * @param[in] idx + * mlx5 flow counter index in the container. + * @param[out] ppool + * mlx5 flow counter pool in the container, * * @return - * pointer to flow counter on success, NULL otherwise and rte_errno is set. + * Pointer to the counter, NULL otherwise. */ static struct mlx5_flow_counter * -flow_dv_counter_alloc_fallback(struct rte_eth_dev *dev, uint32_t shared, - uint32_t id) -{ - struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_flow_counter *cnt = NULL; - struct mlx5_devx_obj *dcs = NULL; - - if (!priv->config.devx) { - rte_errno = ENOTSUP; - return NULL; - } - if (shared) { - TAILQ_FOREACH(cnt, &priv->sh->cmng.flow_counters, next) { - if (cnt->shared && cnt->id == id) { - cnt->ref_cnt++; - return cnt; - } - } - } - dcs = mlx5_devx_cmd_flow_counter_alloc(priv->sh->ctx, 0); - if (!dcs) - return NULL; - cnt = rte_calloc(__func__, 1, sizeof(*cnt), 0); - if (!cnt) { - claim_zero(mlx5_devx_cmd_destroy(cnt->dcs)); - rte_errno = ENOMEM; - return NULL; - } - struct mlx5_flow_counter tmpl = { - .shared = shared, - .ref_cnt = 1, - .id = id, - .dcs = dcs, - }; - tmpl.action = mlx5_glue->dv_create_flow_action_counter(dcs->obj, 0); - if (!tmpl.action) { - claim_zero(mlx5_devx_cmd_destroy(cnt->dcs)); - rte_errno = errno; - rte_free(cnt); - return NULL; - } - *cnt = tmpl; - TAILQ_INSERT_HEAD(&priv->sh->cmng.flow_counters, cnt, next); - return cnt; -} - -/** - * Release a flow counter. - * - * @param[in] dev - * Pointer to the Ethernet device structure. - * @param[in] counter - * Pointer to the counter handler. - */ -static void -flow_dv_counter_release_fallback(struct rte_eth_dev *dev, - struct mlx5_flow_counter *counter) +flow_dv_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_pools_container *cont; + struct mlx5_flow_counter_pool *pool; + uint32_t batch = 0, age = 0; - if (!counter) - return; - if (--counter->ref_cnt == 0) { - TAILQ_REMOVE(&priv->sh->cmng.flow_counters, counter, next); - claim_zero(mlx5_devx_cmd_destroy(counter->dcs)); - rte_free(counter); - } -} - -/** - * Query a devx flow counter. - * - * @param[in] dev - * Pointer to the Ethernet device structure. - * @param[in] cnt - * Pointer to the flow counter. - * @param[out] pkts - * The statistics value of packets. - * @param[out] bytes - * The statistics value of bytes. - * - * @return - * 0 on success, otherwise a negative errno value and rte_errno is set. - */ -static inline int -_flow_dv_query_count_fallback(struct rte_eth_dev *dev __rte_unused, - struct mlx5_flow_counter *cnt, uint64_t *pkts, - uint64_t *bytes) -{ - return mlx5_devx_cmd_flow_counter_query(cnt->dcs, 0, 0, pkts, bytes, - 0, NULL, NULL, 0); -} - -/** - * Get a pool by a counter. - * - * @param[in] cnt - * Pointer to the counter. - * - * @return - * The counter pool. - */ -static struct mlx5_flow_counter_pool * -flow_dv_counter_pool_get(struct mlx5_flow_counter *cnt) -{ - if (!cnt->batch) { - cnt -= cnt->dcs->id % MLX5_COUNTERS_PER_POOL; - return (struct mlx5_flow_counter_pool *)cnt - 1; + idx--; + age = MLX_CNT_IS_AGE(idx); + idx = age ? idx - MLX5_CNT_AGE_OFFSET : idx; + if (idx >= MLX5_CNT_BATCH_OFFSET) { + idx -= MLX5_CNT_BATCH_OFFSET; + batch = 1; } - return cnt->pool; + cont = MLX5_CNT_CONTAINER(priv->sh, batch, 0, age); + MLX5_ASSERT(idx / MLX5_COUNTERS_PER_POOL < cont->n); + pool = cont->pools[idx / MLX5_COUNTERS_PER_POOL]; + MLX5_ASSERT(pool); + if (ppool) + *ppool = pool; + return MLX5_POOL_GET_CNT(pool, idx % MLX5_COUNTERS_PER_POOL); } /** @@ -3873,15 +4061,26 @@ flow_dv_counter_pool_get(struct mlx5_flow_counter *cnt) static struct mlx5_flow_counter_pool * flow_dv_find_pool_by_id(struct mlx5_pools_container *cont, int id) { - struct mlx5_flow_counter_pool *pool; + uint32_t i; + uint32_t n_valid = rte_atomic16_read(&cont->n_valid); - TAILQ_FOREACH(pool, &cont->pool_list, next) { + for (i = 0; i < n_valid; i++) { + struct mlx5_flow_counter_pool *pool = cont->pools[i]; int base = (pool->min_dcs->id / MLX5_COUNTERS_PER_POOL) * - MLX5_COUNTERS_PER_POOL; + MLX5_COUNTERS_PER_POOL; - if (id >= base && id < base + MLX5_COUNTERS_PER_POOL) + if (id >= base && id < base + MLX5_COUNTERS_PER_POOL) { + /* + * Move the pool to the head, as counter allocate + * always gets the first pool in the container. + */ + if (pool != TAILQ_FIRST(&cont->pool_list)) { + TAILQ_REMOVE(&cont->pool_list, pool, next); + TAILQ_INSERT_HEAD(&cont->pool_list, pool, next); + } return pool; - }; + } + } return NULL; } @@ -3901,8 +4100,8 @@ flow_dv_find_pool_by_id(struct mlx5_pools_container *cont, int id) static struct mlx5_counter_stats_mem_mng * flow_dv_create_counter_stat_mem_mng(struct rte_eth_dev *dev, int raws_n) { - struct mlx5_ibv_shared *sh = ((struct mlx5_priv *) - (dev->data->dev_private))->sh; + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_ibv_shared *sh = priv->sh; struct mlx5_devx_mkey_attr mkey_attr; struct mlx5_counter_stats_mem_mng *mem_mng; volatile struct flow_counter_stats *raw_data; @@ -3930,6 +4129,14 @@ flow_dv_create_counter_stat_mem_mng(struct rte_eth_dev *dev, int raws_n) mkey_attr.size = size; mkey_attr.umem_id = mem_mng->umem->umem_id; mkey_attr.pd = sh->pdn; + mkey_attr.log_entity_size = 0; + mkey_attr.pg_access = 0; + mkey_attr.klm_array = NULL; + mkey_attr.klm_num = 0; + if (priv->config.hca_attr.relaxed_ordering_write && + priv->config.hca_attr.relaxed_ordering_read && + !haswell_broadwell_cpu) + mkey_attr.relaxed_ordering = 1; mem_mng->dm = mlx5_devx_cmd_mkey_create(sh->ctx, &mkey_attr); if (!mem_mng->dm) { mlx5_glue->devx_umem_dereg(mem_mng->umem); @@ -3954,24 +4161,29 @@ flow_dv_create_counter_stat_mem_mng(struct rte_eth_dev *dev, int raws_n) * Pointer to the Ethernet device structure. * @param[in] batch * Whether the pool is for counter that was allocated by batch command. + * @param[in] age + * Whether the pool is for Aging counter. * * @return * The new container pointer on success, otherwise NULL and rte_errno is set. */ static struct mlx5_pools_container * -flow_dv_container_resize(struct rte_eth_dev *dev, uint32_t batch) +flow_dv_container_resize(struct rte_eth_dev *dev, + uint32_t batch, uint32_t age) { struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_pools_container *cont = - MLX5_CNT_CONTAINER(priv->sh, batch, 0); + MLX5_CNT_CONTAINER(priv->sh, batch, 0, age); struct mlx5_pools_container *new_cont = - MLX5_CNT_CONTAINER_UNUSED(priv->sh, batch, 0); - struct mlx5_counter_stats_mem_mng *mem_mng; + MLX5_CNT_CONTAINER_UNUSED(priv->sh, batch, 0, age); + struct mlx5_counter_stats_mem_mng *mem_mng = NULL; uint32_t resize = cont->n + MLX5_CNT_CONTAINER_RESIZE; uint32_t mem_size = sizeof(struct mlx5_flow_counter_pool *) * resize; int i; - if (cont != MLX5_CNT_CONTAINER(priv->sh, batch, 1)) { + /* Fallback mode has no background thread. Skip the check. */ + if (!priv->counter_fallback && + cont != MLX5_CNT_CONTAINER(priv->sh, batch, 1, age)) { /* The last resize still hasn't detected by the host thread. */ rte_errno = EAGAIN; return NULL; @@ -3984,16 +4196,29 @@ flow_dv_container_resize(struct rte_eth_dev *dev, uint32_t batch) if (cont->n) memcpy(new_cont->pools, cont->pools, cont->n * sizeof(struct mlx5_flow_counter_pool *)); - mem_mng = flow_dv_create_counter_stat_mem_mng(dev, - MLX5_CNT_CONTAINER_RESIZE + MLX5_MAX_PENDING_QUERIES); - if (!mem_mng) { - rte_free(new_cont->pools); - return NULL; + /* + * Fallback mode query the counter directly, no background query + * resources are needed. + */ + if (!priv->counter_fallback) { + mem_mng = flow_dv_create_counter_stat_mem_mng(dev, + MLX5_CNT_CONTAINER_RESIZE + MLX5_MAX_PENDING_QUERIES); + if (!mem_mng) { + rte_free(new_cont->pools); + return NULL; + } + for (i = 0; i < MLX5_MAX_PENDING_QUERIES; ++i) + LIST_INSERT_HEAD(&priv->sh->cmng.free_stat_raws, + mem_mng->raws + + MLX5_CNT_CONTAINER_RESIZE + + i, next); + } else { + /* + * Release the old container pools directly as no background + * thread helps that. + */ + rte_free(cont->pools); } - for (i = 0; i < MLX5_MAX_PENDING_QUERIES; ++i) - LIST_INSERT_HEAD(&priv->sh->cmng.free_stat_raws, - mem_mng->raws + MLX5_CNT_CONTAINER_RESIZE + - i, next); new_cont->n = resize; rte_atomic16_set(&new_cont->n_valid, rte_atomic16_read(&cont->n_valid)); TAILQ_INIT(&new_cont->pool_list); @@ -4001,7 +4226,7 @@ flow_dv_container_resize(struct rte_eth_dev *dev, uint32_t batch) new_cont->init_mem_mng = mem_mng; rte_cio_wmb(); /* Flip the master container. */ - priv->sh->cmng.mhi[batch] ^= (uint8_t)1; + priv->sh->cmng.mhi[batch][age] ^= (uint8_t)1; return new_cont; } @@ -4011,7 +4236,7 @@ flow_dv_container_resize(struct rte_eth_dev *dev, uint32_t batch) * @param[in] dev * Pointer to the Ethernet device structure. * @param[in] cnt - * Pointer to the flow counter. + * Index to the flow counter. * @param[out] pkts * The statistics value of packets. * @param[out] bytes @@ -4021,17 +4246,23 @@ flow_dv_container_resize(struct rte_eth_dev *dev, uint32_t batch) * 0 on success, otherwise a negative errno value and rte_errno is set. */ static inline int -_flow_dv_query_count(struct rte_eth_dev *dev, - struct mlx5_flow_counter *cnt, uint64_t *pkts, +_flow_dv_query_count(struct rte_eth_dev *dev, uint32_t counter, uint64_t *pkts, uint64_t *bytes) { struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_flow_counter_pool *pool = - flow_dv_counter_pool_get(cnt); - int offset = cnt - &pool->counters_raw[0]; + struct mlx5_flow_counter_pool *pool = NULL; + struct mlx5_flow_counter *cnt; + struct mlx5_flow_counter_ext *cnt_ext = NULL; + int offset; - if (priv->counter_fallback) - return _flow_dv_query_count_fallback(dev, cnt, pkts, bytes); + cnt = flow_dv_counter_get_by_idx(dev, counter, &pool); + MLX5_ASSERT(pool); + if (counter < MLX5_CNT_BATCH_OFFSET) { + cnt_ext = MLX5_CNT_TO_CNT_EXT(pool, cnt); + if (priv->counter_fallback) + return mlx5_devx_cmd_flow_counter_query(cnt_ext->dcs, 0, + 0, pkts, bytes, 0, NULL, NULL, 0); + } rte_spinlock_lock(&pool->sl); /* @@ -4039,10 +4270,11 @@ _flow_dv_query_count(struct rte_eth_dev *dev, * current allocated in parallel to the host reading. * In this case the new counter values must be reported as 0. */ - if (unlikely(!cnt->batch && cnt->dcs->id < pool->raw->min_dcs_id)) { + if (unlikely(cnt_ext && cnt_ext->dcs->id < pool->raw->min_dcs_id)) { *pkts = 0; *bytes = 0; } else { + offset = MLX5_CNT_ARRAY_IDX(pool, cnt); *pkts = rte_be_to_cpu_64(pool->raw->data[offset].hits); *bytes = rte_be_to_cpu_64(pool->raw->data[offset].bytes); } @@ -4059,52 +4291,105 @@ _flow_dv_query_count(struct rte_eth_dev *dev, * The devX counter handle. * @param[in] batch * Whether the pool is for counter that was allocated by batch command. + * @param[in] age + * Whether the pool is for counter that was allocated for aging. + * @param[in/out] cont_cur + * Pointer to the container pointer, it will be update in pool resize. * * @return - * A new pool pointer on success, NULL otherwise and rte_errno is set. + * The pool container pointer on success, NULL otherwise and rte_errno is set. */ -static struct mlx5_flow_counter_pool * +static struct mlx5_pools_container * flow_dv_pool_create(struct rte_eth_dev *dev, struct mlx5_devx_obj *dcs, - uint32_t batch) + uint32_t batch, uint32_t age) { struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_flow_counter_pool *pool; struct mlx5_pools_container *cont = MLX5_CNT_CONTAINER(priv->sh, batch, - 0); + 0, age); int16_t n_valid = rte_atomic16_read(&cont->n_valid); - uint32_t size; + uint32_t size = sizeof(*pool); if (cont->n == n_valid) { - cont = flow_dv_container_resize(dev, batch); + cont = flow_dv_container_resize(dev, batch, age); if (!cont) return NULL; } - size = sizeof(*pool) + MLX5_COUNTERS_PER_POOL * - sizeof(struct mlx5_flow_counter); + size += MLX5_COUNTERS_PER_POOL * CNT_SIZE; + size += (batch ? 0 : MLX5_COUNTERS_PER_POOL * CNTEXT_SIZE); + size += (!age ? 0 : MLX5_COUNTERS_PER_POOL * AGE_SIZE); pool = rte_calloc(__func__, 1, size, 0); if (!pool) { rte_errno = ENOMEM; return NULL; } pool->min_dcs = dcs; - pool->raw = cont->init_mem_mng->raws + n_valid % + if (!priv->counter_fallback) + pool->raw = cont->init_mem_mng->raws + n_valid % MLX5_CNT_CONTAINER_RESIZE; pool->raw_hw = NULL; + pool->type = 0; + pool->type |= (batch ? 0 : CNT_POOL_TYPE_EXT); + pool->type |= (!age ? 0 : CNT_POOL_TYPE_AGE); rte_spinlock_init(&pool->sl); /* * The generation of the new allocated counters in this pool is 0, 2 in * the pool generation makes all the counters valid for allocation. + * The start and end query generation protect the counters be released + * between the query and update gap period will not be reallocated + * without the last query finished and stats updated to the memory. + */ + rte_atomic64_set(&pool->start_query_gen, 0x2); + /* + * There's no background query thread for fallback mode, set the + * end_query_gen to the maximum value since no need to wait for + * statistics update. */ - rte_atomic64_set(&pool->query_gen, 0x2); + rte_atomic64_set(&pool->end_query_gen, priv->counter_fallback ? + INT64_MAX : 0x2); TAILQ_INIT(&pool->counters); - TAILQ_INSERT_TAIL(&cont->pool_list, pool, next); + TAILQ_INSERT_HEAD(&cont->pool_list, pool, next); + pool->index = n_valid; cont->pools[n_valid] = pool; /* Pool initialization must be updated before host thread access. */ rte_cio_wmb(); rte_atomic16_add(&cont->n_valid, 1); - return pool; + return cont; } +/** + * Update the minimum dcs-id for aged or no-aged counter pool. + * + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] pool + * Current counter pool. + * @param[in] batch + * Whether the pool is for counter that was allocated by batch command. + * @param[in] age + * Whether the counter is for aging. + */ +static void +flow_dv_counter_update_min_dcs(struct rte_eth_dev *dev, + struct mlx5_flow_counter_pool *pool, + uint32_t batch, uint32_t age) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_counter_pool *other; + struct mlx5_pools_container *cont; + + cont = MLX5_CNT_CONTAINER(priv->sh, batch, 0, (age ^ 0x1)); + other = flow_dv_find_pool_by_id(cont, pool->min_dcs->id); + if (!other) + return; + if (pool->min_dcs->id < other->min_dcs->id) { + rte_atomic64_set(&other->a64_dcs, + rte_atomic64_read(&pool->a64_dcs)); + } else { + rte_atomic64_set(&pool->a64_dcs, + rte_atomic64_read(&other->a64_dcs)); + } +} /** * Prepare a new counter and/or a new counter pool. * @@ -4114,44 +4399,51 @@ flow_dv_pool_create(struct rte_eth_dev *dev, struct mlx5_devx_obj *dcs, * Where to put the pointer of a new counter. * @param[in] batch * Whether the pool is for counter that was allocated by batch command. + * @param[in] age + * Whether the pool is for counter that was allocated for aging. * * @return - * The free counter pool pointer and @p cnt_free is set on success, + * The counter container pointer and @p cnt_free is set on success, * NULL otherwise and rte_errno is set. */ -static struct mlx5_flow_counter_pool * +static struct mlx5_pools_container * flow_dv_counter_pool_prepare(struct rte_eth_dev *dev, struct mlx5_flow_counter **cnt_free, - uint32_t batch) + uint32_t batch, uint32_t age) { struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_pools_container *cont; struct mlx5_flow_counter_pool *pool; struct mlx5_devx_obj *dcs = NULL; struct mlx5_flow_counter *cnt; uint32_t i; + cont = MLX5_CNT_CONTAINER(priv->sh, batch, 0, age); if (!batch) { /* bulk_bitmap must be 0 for single counter allocation. */ dcs = mlx5_devx_cmd_flow_counter_alloc(priv->sh->ctx, 0); if (!dcs) return NULL; - pool = flow_dv_find_pool_by_id - (MLX5_CNT_CONTAINER(priv->sh, batch, 0), dcs->id); + pool = flow_dv_find_pool_by_id(cont, dcs->id); if (!pool) { - pool = flow_dv_pool_create(dev, dcs, batch); - if (!pool) { + cont = flow_dv_pool_create(dev, dcs, batch, age); + if (!cont) { mlx5_devx_cmd_destroy(dcs); return NULL; } + pool = TAILQ_FIRST(&cont->pool_list); } else if (dcs->id < pool->min_dcs->id) { rte_atomic64_set(&pool->a64_dcs, (int64_t)(uintptr_t)dcs); } - cnt = &pool->counters_raw[dcs->id % MLX5_COUNTERS_PER_POOL]; + flow_dv_counter_update_min_dcs(dev, + pool, batch, age); + i = dcs->id % MLX5_COUNTERS_PER_POOL; + cnt = MLX5_POOL_GET_CNT(pool, i); TAILQ_INSERT_HEAD(&pool->counters, cnt, next); - cnt->dcs = dcs; + MLX5_GET_POOL_CNT_EXT(pool, i)->dcs = dcs; *cnt_free = cnt; - return pool; + return cont; } /* bulk_bitmap is in 128 counters units. */ if (priv->config.hca_attr.flow_counter_bulk_alloc_bitmap & 0x4) @@ -4160,18 +4452,18 @@ flow_dv_counter_pool_prepare(struct rte_eth_dev *dev, rte_errno = ENODATA; return NULL; } - pool = flow_dv_pool_create(dev, dcs, batch); - if (!pool) { + cont = flow_dv_pool_create(dev, dcs, batch, age); + if (!cont) { mlx5_devx_cmd_destroy(dcs); return NULL; } + pool = TAILQ_FIRST(&cont->pool_list); for (i = 0; i < MLX5_COUNTERS_PER_POOL; ++i) { - cnt = &pool->counters_raw[i]; - cnt->pool = pool; + cnt = MLX5_POOL_GET_CNT(pool, i); TAILQ_INSERT_HEAD(&pool->counters, cnt, next); } - *cnt_free = &pool->counters_raw[0]; - return pool; + *cnt_free = MLX5_POOL_GET_CNT(pool, 0); + return cont; } /** @@ -4181,23 +4473,30 @@ flow_dv_counter_pool_prepare(struct rte_eth_dev *dev, * Pointer to the relevant counter pool container. * @param[in] id * The shared counter ID to search. + * @param[out] ppool + * mlx5 flow counter pool in the container, * * @return - * NULL if not existed, otherwise pointer to the shared counter. + * NULL if not existed, otherwise pointer to the shared extend counter. */ -static struct mlx5_flow_counter * -flow_dv_counter_shared_search(struct mlx5_pools_container *cont, - uint32_t id) +static struct mlx5_flow_counter_ext * +flow_dv_counter_shared_search(struct mlx5_pools_container *cont, uint32_t id, + struct mlx5_flow_counter_pool **ppool) { - static struct mlx5_flow_counter *cnt; + static struct mlx5_flow_counter_ext *cnt; struct mlx5_flow_counter_pool *pool; - int i; + uint32_t i; + uint32_t n_valid = rte_atomic16_read(&cont->n_valid); - TAILQ_FOREACH(pool, &cont->pool_list, next) { + for (i = 0; i < n_valid; i++) { + pool = cont->pools[i]; for (i = 0; i < MLX5_COUNTERS_PER_POOL; ++i) { - cnt = &pool->counters_raw[i]; - if (cnt->ref_cnt && cnt->shared && cnt->id == id) + cnt = MLX5_GET_POOL_CNT_EXT(pool, i); + if (cnt->ref_cnt && cnt->shared && cnt->id == id) { + if (ppool) + *ppool = cont->pools[i]; return cnt; + } } } return NULL; @@ -4214,17 +4513,20 @@ flow_dv_counter_shared_search(struct mlx5_pools_container *cont, * Counter identifier. * @param[in] group * Counter flow group. + * @param[in] age + * Whether the counter was allocated for aging. * * @return - * pointer to flow counter on success, NULL otherwise and rte_errno is set. + * Index to flow counter on success, 0 otherwise and rte_errno is set. */ -static struct mlx5_flow_counter * +static uint32_t flow_dv_counter_alloc(struct rte_eth_dev *dev, uint32_t shared, uint32_t id, - uint16_t group) + uint16_t group, uint32_t age) { struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_flow_counter_pool *pool = NULL; struct mlx5_flow_counter *cnt_free = NULL; + struct mlx5_flow_counter_ext *cnt_ext = NULL; /* * Currently group 0 flow counter cannot be assigned to a flow if it is * not the first one in the batch counter allocation, so it is better @@ -4233,25 +4535,27 @@ flow_dv_counter_alloc(struct rte_eth_dev *dev, uint32_t shared, uint32_t id, * A counter can be shared between different groups so need to take * shared counters from the single container. */ - uint32_t batch = (group && !shared) ? 1 : 0; + uint32_t batch = (group && !shared && !priv->counter_fallback) ? 1 : 0; struct mlx5_pools_container *cont = MLX5_CNT_CONTAINER(priv->sh, batch, - 0); + 0, age); + uint32_t cnt_idx; - if (priv->counter_fallback) - return flow_dv_counter_alloc_fallback(dev, shared, id); if (!priv->config.devx) { rte_errno = ENOTSUP; - return NULL; + return 0; } if (shared) { - cnt_free = flow_dv_counter_shared_search(cont, id); - if (cnt_free) { - if (cnt_free->ref_cnt + 1 == 0) { + cnt_ext = flow_dv_counter_shared_search(cont, id, &pool); + if (cnt_ext) { + if (cnt_ext->ref_cnt + 1 == 0) { rte_errno = E2BIG; - return NULL; + return 0; } - cnt_free->ref_cnt++; - return cnt_free; + cnt_ext->ref_cnt++; + cnt_idx = pool->index * MLX5_COUNTERS_PER_POOL + + (cnt_ext->dcs->id % MLX5_COUNTERS_PER_POOL) + + 1; + return cnt_idx; } } /* Pools which has a free counters are in the start. */ @@ -4266,44 +4570,52 @@ flow_dv_counter_alloc(struct rte_eth_dev *dev, uint32_t shared, uint32_t id, * updated too. */ cnt_free = TAILQ_FIRST(&pool->counters); - if (cnt_free && cnt_free->query_gen + 1 < - rte_atomic64_read(&pool->query_gen)) + if (cnt_free && cnt_free->query_gen < + rte_atomic64_read(&pool->end_query_gen)) break; cnt_free = NULL; } if (!cnt_free) { - pool = flow_dv_counter_pool_prepare(dev, &cnt_free, batch); - if (!pool) - return NULL; + cont = flow_dv_counter_pool_prepare(dev, &cnt_free, batch, age); + if (!cont) + return 0; + pool = TAILQ_FIRST(&cont->pool_list); } - cnt_free->batch = batch; + if (!batch) + cnt_ext = MLX5_CNT_TO_CNT_EXT(pool, cnt_free); /* Create a DV counter action only in the first time usage. */ if (!cnt_free->action) { uint16_t offset; struct mlx5_devx_obj *dcs; if (batch) { - offset = cnt_free - &pool->counters_raw[0]; + offset = MLX5_CNT_ARRAY_IDX(pool, cnt_free); dcs = pool->min_dcs; } else { offset = 0; - dcs = cnt_free->dcs; + dcs = cnt_ext->dcs; } cnt_free->action = mlx5_glue->dv_create_flow_action_counter (dcs->obj, offset); if (!cnt_free->action) { rte_errno = errno; - return NULL; + return 0; } } + cnt_idx = MLX5_MAKE_CNT_IDX(pool->index, + MLX5_CNT_ARRAY_IDX(pool, cnt_free)); + cnt_idx += batch * MLX5_CNT_BATCH_OFFSET; + cnt_idx += age * MLX5_CNT_AGE_OFFSET; /* Update the counter reset values. */ - if (_flow_dv_query_count(dev, cnt_free, &cnt_free->hits, + if (_flow_dv_query_count(dev, cnt_idx, &cnt_free->hits, &cnt_free->bytes)) - return NULL; - cnt_free->shared = shared; - cnt_free->ref_cnt = 1; - cnt_free->id = id; - if (!priv->sh->cmng.query_thread_on) + return 0; + if (cnt_ext) { + cnt_ext->shared = shared; + cnt_ext->ref_cnt = 1; + cnt_ext->id = id; + } + if (!priv->counter_fallback && !priv->sh->cmng.query_thread_on) /* Start the asynchronous batch query by the host thread. */ mlx5_set_query_alarm(priv->sh); TAILQ_REMOVE(&pool->counters, cnt_free, next); @@ -4312,37 +4624,102 @@ flow_dv_counter_alloc(struct rte_eth_dev *dev, uint32_t shared, uint32_t id, TAILQ_REMOVE(&cont->pool_list, pool, next); TAILQ_INSERT_TAIL(&cont->pool_list, pool, next); } - return cnt_free; + return cnt_idx; } /** - * Release a flow counter. + * Get age param from counter index. + * + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] counter + * Index to the counter handler. + * + * @return + * The aging parameter specified for the counter index. + */ +static struct mlx5_age_param* +flow_dv_counter_idx_get_age(struct rte_eth_dev *dev, + uint32_t counter) +{ + struct mlx5_flow_counter *cnt; + struct mlx5_flow_counter_pool *pool = NULL; + + flow_dv_counter_get_by_idx(dev, counter, &pool); + counter = (counter - 1) % MLX5_COUNTERS_PER_POOL; + cnt = MLX5_POOL_GET_CNT(pool, counter); + return MLX5_CNT_TO_AGE(cnt); +} + +/** + * Remove a flow counter from aged counter list. * * @param[in] dev * Pointer to the Ethernet device structure. * @param[in] counter + * Index to the counter handler. + * @param[in] cnt * Pointer to the counter handler. */ static void -flow_dv_counter_release(struct rte_eth_dev *dev, - struct mlx5_flow_counter *counter) +flow_dv_counter_remove_from_age(struct rte_eth_dev *dev, + uint32_t counter, struct mlx5_flow_counter *cnt) { + struct mlx5_age_info *age_info; + struct mlx5_age_param *age_param; struct mlx5_priv *priv = dev->data->dev_private; - if (!counter) - return; - if (priv->counter_fallback) { - flow_dv_counter_release_fallback(dev, counter); - return; + age_info = GET_PORT_AGE_INFO(priv); + age_param = flow_dv_counter_idx_get_age(dev, counter); + if (rte_atomic16_cmpset((volatile uint16_t *) + &age_param->state, + AGE_CANDIDATE, AGE_FREE) + != AGE_CANDIDATE) { + /** + * We need the lock even it is age timeout, + * since counter may still in process. + */ + rte_spinlock_lock(&age_info->aged_sl); + TAILQ_REMOVE(&age_info->aged_counters, cnt, next); + rte_spinlock_unlock(&age_info->aged_sl); } - if (--counter->ref_cnt == 0) { - struct mlx5_flow_counter_pool *pool = - flow_dv_counter_pool_get(counter); + rte_atomic16_set(&age_param->state, AGE_FREE); +} +/** + * Release a flow counter. + * + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] counter + * Index to the counter handler. + */ +static void +flow_dv_counter_release(struct rte_eth_dev *dev, uint32_t counter) +{ + struct mlx5_flow_counter_pool *pool = NULL; + struct mlx5_flow_counter *cnt; + struct mlx5_flow_counter_ext *cnt_ext = NULL; - /* Put the counter in the end - the last updated one. */ - TAILQ_INSERT_TAIL(&pool->counters, counter, next); - counter->query_gen = rte_atomic64_read(&pool->query_gen); + if (!counter) + return; + cnt = flow_dv_counter_get_by_idx(dev, counter, &pool); + MLX5_ASSERT(pool); + if (counter < MLX5_CNT_BATCH_OFFSET) { + cnt_ext = MLX5_CNT_TO_CNT_EXT(pool, cnt); + if (cnt_ext && --cnt_ext->ref_cnt) + return; } + if (IS_AGE_POOL(pool)) + flow_dv_counter_remove_from_age(dev, counter, cnt); + /* Put the counter in the end - the last updated one. */ + TAILQ_INSERT_TAIL(&pool->counters, cnt, next); + /* + * Counters released between query trigger and handler need + * to wait the next round of query. Since the packets arrive + * in the gap period will not be taken into account to the + * old counter. + */ + cnt->query_gen = rte_atomic64_read(&pool->start_query_gen); } /** @@ -4359,7 +4736,9 @@ flow_dv_counter_release(struct rte_eth_dev *dev, * Pointer to error structure. * * @return - * 0 on success, a negative errno value otherwise and rte_errno is set. + * - 0 on success and non root table. + * - 1 on success and root table. + * - a negative errno value otherwise and rte_errno is set. */ static int flow_dv_validate_attributes(struct rte_eth_dev *dev, @@ -4369,6 +4748,7 @@ flow_dv_validate_attributes(struct rte_eth_dev *dev, { struct mlx5_priv *priv = dev->data->dev_private; uint32_t priority_max = priv->config.flow_prio - 1; + int ret = 0; #ifndef HAVE_MLX5DV_DR if (attributes->group) @@ -4377,14 +4757,15 @@ flow_dv_validate_attributes(struct rte_eth_dev *dev, NULL, "groups are not supported"); #else - uint32_t table; - int ret; + uint32_t table = 0; ret = mlx5_flow_group_to_table(attributes, external, - attributes->group, + attributes->group, !!priv->fdb_def_rule, &table, error); if (ret) return ret; + if (!table) + ret = MLX5DV_DR_ACTION_FLAGS_ROOT_LEVEL; #endif if (attributes->priority != MLX5_FLOW_PRIO_RSVD && attributes->priority >= priority_max) @@ -4414,7 +4795,7 @@ flow_dv_validate_attributes(struct rte_eth_dev *dev, RTE_FLOW_ERROR_TYPE_ATTR, NULL, "must specify exactly one of " "ingress or egress"); - return 0; + return ret; } /** @@ -4430,6 +4811,8 @@ flow_dv_validate_attributes(struct rte_eth_dev *dev, * 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. * @@ -4440,7 +4823,7 @@ static int flow_dv_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, struct rte_flow_error *error) + bool external, int hairpin, struct rte_flow_error *error) { int ret; uint64_t action_flags = 0; @@ -4449,22 +4832,53 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, uint8_t next_protocol = 0xff; uint16_t ether_type = 0; int actions_n = 0; + uint8_t item_ipv6_proto = 0; const struct rte_flow_item *gre_item = NULL; - struct rte_flow_item_tcp nic_tcp_mask = { + const struct rte_flow_action_raw_decap *decap; + const struct rte_flow_action_raw_encap *encap; + const struct rte_flow_action_rss *rss; + const struct rte_flow_item_tcp nic_tcp_mask = { .hdr = { .tcp_flags = 0xFF, .src_port = RTE_BE16(UINT16_MAX), .dst_port = RTE_BE16(UINT16_MAX), } }; + const struct rte_flow_item_ipv4 nic_ipv4_mask = { + .hdr = { + .src_addr = RTE_BE32(0xffffffff), + .dst_addr = RTE_BE32(0xffffffff), + .type_of_service = 0xff, + .next_proto_id = 0xff, + .time_to_live = 0xff, + }, + }; + const struct rte_flow_item_ipv6 nic_ipv6_mask = { + .hdr = { + .src_addr = + "\xff\xff\xff\xff\xff\xff\xff\xff" + "\xff\xff\xff\xff\xff\xff\xff\xff", + .dst_addr = + "\xff\xff\xff\xff\xff\xff\xff\xff" + "\xff\xff\xff\xff\xff\xff\xff\xff", + .vtc_flow = RTE_BE32(0xffffffff), + .proto = 0xff, + .hop_limits = 0xff, + }, + }; struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_dev_config *dev_conf = &priv->config; + uint16_t queue_index = 0xFFFF; + const struct rte_flow_item_vlan *vlan_m = NULL; + int16_t rw_act_num = 0; + uint64_t is_root; if (items == NULL) return -1; ret = flow_dv_validate_attributes(dev, attr, external, error); if (ret < 0) return ret; + is_root = (uint64_t)ret; for (; items->type != RTE_FLOW_ITEM_TYPE_END; items++) { int tunnel = !!(item_flags & MLX5_FLOW_LAYER_TUNNEL); int type = items->type; @@ -4499,8 +4913,8 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, } break; case RTE_FLOW_ITEM_TYPE_VLAN: - ret = mlx5_flow_validate_item_vlan(items, item_flags, - dev, error); + ret = flow_dv_validate_item_vlan(items, item_flags, + dev, error); if (ret < 0) return ret; last_item = tunnel ? MLX5_FLOW_LAYER_INNER_VLAN : @@ -4516,13 +4930,17 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, } else { ether_type = 0; } + /* Store outer VLAN mask for of_push_vlan action. */ + if (!tunnel) + vlan_m = items->mask; break; case RTE_FLOW_ITEM_TYPE_IPV4: mlx5_flow_tunnel_ip_check(items, next_protocol, &item_flags, &tunnel); ret = mlx5_flow_validate_item_ipv4(items, item_flags, last_item, - ether_type, NULL, + ether_type, + &nic_ipv4_mask, error); if (ret < 0) return ret; @@ -4547,7 +4965,8 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, &item_flags, &tunnel); ret = mlx5_flow_validate_item_ipv6(items, item_flags, last_item, - ether_type, NULL, + ether_type, + &nic_ipv6_mask, error); if (ret < 0) return ret; @@ -4556,6 +4975,9 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, if (items->mask != NULL && ((const struct rte_flow_item_ipv6 *) items->mask)->hdr.proto) { + item_ipv6_proto = + ((const struct rte_flow_item_ipv6 *) + items->spec)->hdr.proto; next_protocol = ((const struct rte_flow_item_ipv6 *) items->spec)->hdr.proto; @@ -4670,6 +5092,7 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, error); if (ret < 0) return ret; + item_ipv6_proto = IPPROTO_ICMPV6; last_item = MLX5_FLOW_LAYER_ICMP6; break; case RTE_FLOW_ITEM_TYPE_TAG: @@ -4732,6 +5155,7 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, action_flags |= MLX5_FLOW_ACTION_FLAG; ++actions_n; } + rw_act_num += MLX5_ACT_NUM_SET_MARK; break; case RTE_FLOW_ACTION_TYPE_MARK: ret = flow_dv_validate_action_mark(dev, actions, @@ -4750,6 +5174,7 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, action_flags |= MLX5_FLOW_ACTION_MARK; ++actions_n; } + rw_act_num += MLX5_ACT_NUM_SET_MARK; break; case RTE_FLOW_ACTION_TYPE_SET_META: ret = flow_dv_validate_action_set_meta(dev, actions, @@ -4761,6 +5186,7 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, if (!(action_flags & MLX5_FLOW_MODIFY_HDR_ACTIONS)) ++actions_n; action_flags |= MLX5_FLOW_ACTION_SET_META; + rw_act_num += MLX5_ACT_NUM_SET_META; break; case RTE_FLOW_ACTION_TYPE_SET_TAG: ret = flow_dv_validate_action_set_tag(dev, actions, @@ -4772,6 +5198,7 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, if (!(action_flags & MLX5_FLOW_MODIFY_HDR_ACTIONS)) ++actions_n; action_flags |= MLX5_FLOW_ACTION_SET_TAG; + rw_act_num += MLX5_ACT_NUM_SET_TAG; break; case RTE_FLOW_ACTION_TYPE_DROP: ret = mlx5_flow_validate_action_drop(action_flags, @@ -4787,16 +5214,21 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, attr, error); if (ret < 0) return ret; + queue_index = ((const struct rte_flow_action_queue *) + (actions->conf))->index; action_flags |= MLX5_FLOW_ACTION_QUEUE; ++actions_n; break; case RTE_FLOW_ACTION_TYPE_RSS: + rss = actions->conf; ret = mlx5_flow_validate_action_rss(actions, action_flags, dev, attr, item_flags, error); if (ret < 0) return ret; + if (rss != NULL && rss->queue_num) + queue_index = rss->queue[0]; action_flags |= MLX5_FLOW_ACTION_RSS; ++actions_n; break; @@ -4818,8 +5250,9 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, ++actions_n; break; case RTE_FLOW_ACTION_TYPE_OF_PUSH_VLAN: - ret = flow_dv_validate_action_push_vlan(action_flags, - item_flags, + ret = flow_dv_validate_action_push_vlan(dev, + action_flags, + vlan_m, actions, attr, error); if (ret < 0) @@ -4843,49 +5276,52 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, return ret; /* Count VID with push_vlan command. */ action_flags |= MLX5_FLOW_ACTION_OF_SET_VLAN_VID; + rw_act_num += MLX5_ACT_NUM_MDF_VID; break; case RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP: case RTE_FLOW_ACTION_TYPE_NVGRE_ENCAP: - ret = flow_dv_validate_action_l2_encap(action_flags, + ret = flow_dv_validate_action_l2_encap(dev, + action_flags, actions, attr, error); if (ret < 0) return ret; - action_flags |= actions->type == - RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP ? - MLX5_FLOW_ACTION_VXLAN_ENCAP : - MLX5_FLOW_ACTION_NVGRE_ENCAP; + action_flags |= MLX5_FLOW_ACTION_ENCAP; ++actions_n; break; case RTE_FLOW_ACTION_TYPE_VXLAN_DECAP: case RTE_FLOW_ACTION_TYPE_NVGRE_DECAP: - ret = flow_dv_validate_action_l2_decap(action_flags, - attr, error); + ret = flow_dv_validate_action_decap(dev, action_flags, + attr, error); if (ret < 0) return ret; - action_flags |= actions->type == - RTE_FLOW_ACTION_TYPE_VXLAN_DECAP ? - MLX5_FLOW_ACTION_VXLAN_DECAP : - MLX5_FLOW_ACTION_NVGRE_DECAP; + action_flags |= MLX5_FLOW_ACTION_DECAP; ++actions_n; break; case RTE_FLOW_ACTION_TYPE_RAW_ENCAP: - ret = flow_dv_validate_action_raw_encap(action_flags, - actions, attr, - error); + ret = flow_dv_validate_action_raw_encap_decap + (dev, NULL, actions->conf, attr, &action_flags, + &actions_n, error); if (ret < 0) return ret; - action_flags |= MLX5_FLOW_ACTION_RAW_ENCAP; - ++actions_n; break; case RTE_FLOW_ACTION_TYPE_RAW_DECAP: - ret = flow_dv_validate_action_raw_decap(action_flags, - actions, attr, - error); + decap = actions->conf; + while ((++actions)->type == RTE_FLOW_ACTION_TYPE_VOID) + ; + if (actions->type != RTE_FLOW_ACTION_TYPE_RAW_ENCAP) { + encap = NULL; + actions--; + } else { + encap = actions->conf; + } + ret = flow_dv_validate_action_raw_encap_decap + (dev, + decap ? decap : &empty_decap, encap, + attr, &action_flags, &actions_n, + error); if (ret < 0) return ret; - action_flags |= MLX5_FLOW_ACTION_RAW_DECAP; - ++actions_n; break; case RTE_FLOW_ACTION_TYPE_SET_MAC_SRC: case RTE_FLOW_ACTION_TYPE_SET_MAC_DST: @@ -4902,8 +5338,15 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, RTE_FLOW_ACTION_TYPE_SET_MAC_SRC ? MLX5_FLOW_ACTION_SET_MAC_SRC : MLX5_FLOW_ACTION_SET_MAC_DST; + /* + * Even if the source and destination MAC addresses have + * overlap in the header with 4B alignment, the convert + * function will handle them separately and 4 SW actions + * will be created. And 2 actions will be added each + * time no matter how many bytes of address will be set. + */ + rw_act_num += MLX5_ACT_NUM_MDF_MAC; break; - case RTE_FLOW_ACTION_TYPE_SET_IPV4_SRC: case RTE_FLOW_ACTION_TYPE_SET_IPV4_DST: ret = flow_dv_validate_action_modify_ipv4(action_flags, @@ -4919,6 +5362,7 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, RTE_FLOW_ACTION_TYPE_SET_IPV4_SRC ? MLX5_FLOW_ACTION_SET_IPV4_SRC : MLX5_FLOW_ACTION_SET_IPV4_DST; + rw_act_num += MLX5_ACT_NUM_MDF_IPV4; break; case RTE_FLOW_ACTION_TYPE_SET_IPV6_SRC: case RTE_FLOW_ACTION_TYPE_SET_IPV6_DST: @@ -4928,6 +5372,12 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, error); if (ret < 0) return ret; + if (item_ipv6_proto == IPPROTO_ICMPV6) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ACTION, + actions, + "Can't change header " + "with ICMPv6 proto"); /* Count all modify-header actions as one action. */ if (!(action_flags & MLX5_FLOW_MODIFY_HDR_ACTIONS)) ++actions_n; @@ -4935,6 +5385,7 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, RTE_FLOW_ACTION_TYPE_SET_IPV6_SRC ? MLX5_FLOW_ACTION_SET_IPV6_SRC : MLX5_FLOW_ACTION_SET_IPV6_DST; + rw_act_num += MLX5_ACT_NUM_MDF_IPV6; break; case RTE_FLOW_ACTION_TYPE_SET_TP_SRC: case RTE_FLOW_ACTION_TYPE_SET_TP_DST: @@ -4951,6 +5402,7 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, RTE_FLOW_ACTION_TYPE_SET_TP_SRC ? MLX5_FLOW_ACTION_SET_TP_SRC : MLX5_FLOW_ACTION_SET_TP_DST; + rw_act_num += MLX5_ACT_NUM_MDF_PORT; break; case RTE_FLOW_ACTION_TYPE_DEC_TTL: case RTE_FLOW_ACTION_TYPE_SET_TTL: @@ -4967,6 +5419,7 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, RTE_FLOW_ACTION_TYPE_SET_TTL ? MLX5_FLOW_ACTION_SET_TTL : MLX5_FLOW_ACTION_DEC_TTL; + rw_act_num += MLX5_ACT_NUM_MDF_TTL; break; case RTE_FLOW_ACTION_TYPE_JUMP: ret = flow_dv_validate_action_jump(actions, @@ -4994,6 +5447,7 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, RTE_FLOW_ACTION_TYPE_INC_TCP_SEQ ? MLX5_FLOW_ACTION_INC_TCP_SEQ : MLX5_FLOW_ACTION_DEC_TCP_SEQ; + rw_act_num += MLX5_ACT_NUM_MDF_TCPSEQ; break; case RTE_FLOW_ACTION_TYPE_INC_TCP_ACK: case RTE_FLOW_ACTION_TYPE_DEC_TCP_ACK: @@ -5011,10 +5465,13 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, RTE_FLOW_ACTION_TYPE_INC_TCP_ACK ? MLX5_FLOW_ACTION_INC_TCP_ACK : MLX5_FLOW_ACTION_DEC_TCP_ACK; + rw_act_num += MLX5_ACT_NUM_MDF_TCPACK; break; - case MLX5_RTE_FLOW_ACTION_TYPE_TAG: case MLX5_RTE_FLOW_ACTION_TYPE_MARK: + break; + case MLX5_RTE_FLOW_ACTION_TYPE_TAG: case MLX5_RTE_FLOW_ACTION_TYPE_COPY_MREG: + rw_act_num += MLX5_ACT_NUM_SET_TAG; break; case RTE_FLOW_ACTION_TYPE_METER: ret = mlx5_flow_validate_action_meter(dev, @@ -5025,6 +5482,17 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, return ret; action_flags |= MLX5_FLOW_ACTION_METER; ++actions_n; + /* Meter action will add one more TAG action. */ + rw_act_num += MLX5_ACT_NUM_SET_TAG; + break; + case RTE_FLOW_ACTION_TYPE_AGE: + ret = flow_dv_validate_action_age(action_flags, + actions, dev, + error); + if (ret < 0) + return ret; + action_flags |= MLX5_FLOW_ACTION_AGE; + ++actions_n; break; case RTE_FLOW_ACTION_TYPE_SET_IPV4_DSCP: ret = flow_dv_validate_action_modify_ipv4_dscp @@ -5038,6 +5506,7 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, if (!(action_flags & MLX5_FLOW_MODIFY_HDR_ACTIONS)) ++actions_n; action_flags |= MLX5_FLOW_ACTION_SET_IPV4_DSCP; + rw_act_num += MLX5_ACT_NUM_SET_DSCP; break; case RTE_FLOW_ACTION_TYPE_SET_IPV6_DSCP: ret = flow_dv_validate_action_modify_ipv6_dscp @@ -5051,6 +5520,7 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, if (!(action_flags & MLX5_FLOW_MODIFY_HDR_ACTIONS)) ++actions_n; action_flags |= MLX5_FLOW_ACTION_SET_IPV6_DSCP; + rw_act_num += MLX5_ACT_NUM_SET_DSCP; break; default: return rte_flow_error_set(error, ENOTSUP, @@ -5059,6 +5529,18 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, "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"); /* Eswitch has few restrictions on using items and actions */ if (attr->transfer) { if (!mlx5_flow_ext_mreg_supported(dev) && @@ -5095,6 +5577,37 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, actions, "no fate action is found"); } + /* Continue validation for Xcap actions.*/ + if ((action_flags & MLX5_FLOW_XCAP_ACTIONS) && (queue_index == 0xFFFF || + mlx5_rxq_get_type(dev, queue_index) != MLX5_RXQ_TYPE_HAIRPIN)) { + if ((action_flags & MLX5_FLOW_XCAP_ACTIONS) == + MLX5_FLOW_XCAP_ACTIONS) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ACTION, + NULL, "encap and decap " + "combination aren't supported"); + if (!attr->transfer && attr->ingress && (action_flags & + MLX5_FLOW_ACTION_ENCAP)) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ACTION, + NULL, "encap is not supported" + " for ingress traffic"); + } + /* Hairpin flow will add one more TAG action. */ + if (hairpin > 0) + rw_act_num += MLX5_ACT_NUM_SET_TAG; + /* extra metadata enabled: one more TAG action will be add. */ + if (dev_conf->dv_flow_en && + dev_conf->dv_xmeta_en != MLX5_XMETA_MODE_LEGACY && + mlx5_flow_ext_mreg_supported(dev)) + rw_act_num += MLX5_ACT_NUM_SET_TAG; + if ((uint32_t)rw_act_num > + flow_dv_modify_hdr_action_max(dev, is_root)) { + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ACTION, + NULL, "too many header modify" + " actions to support"); + } return 0; } @@ -5102,6 +5615,8 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, * Internal preparation function. Allocates the DV flow size, * this size is constant. * + * @param[in] dev + * Pointer to the rte_eth_dev structure. * @param[in] attr * Pointer to the flow attributes. * @param[in] items @@ -5116,28 +5631,49 @@ flow_dv_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, * otherwise NULL and rte_errno is set. */ static struct mlx5_flow * -flow_dv_prepare(const struct rte_flow_attr *attr __rte_unused, +flow_dv_prepare(struct rte_eth_dev *dev, + const struct rte_flow_attr *attr __rte_unused, const struct rte_flow_item items[] __rte_unused, const struct rte_flow_action actions[] __rte_unused, struct rte_flow_error *error) { - size_t size = sizeof(struct mlx5_flow); + uint32_t handle_idx = 0; struct mlx5_flow *dev_flow; + struct mlx5_flow_handle *dev_handle; + struct mlx5_priv *priv = dev->data->dev_private; - dev_flow = rte_calloc(__func__, 1, size, 0); - if (!dev_flow) { + /* In case of corrupting the memory. */ + if (priv->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; } + /* No multi-thread supporting. */ + dev_flow = &((struct mlx5_flow *)priv->inter_flows)[priv->flow_idx++]; + dev_flow->handle = dev_handle; + dev_flow->handle_idx = handle_idx; dev_flow->dv.value.size = MLX5_ST_SZ_BYTES(fte_match_param); + /* + * The matching value needs to be cleared to 0 before using. In the + * past, it will be automatically cleared when using rte_*alloc + * API. The time consumption will be almost the same as before. + */ + memset(dev_flow->dv.value.buf, 0, MLX5_ST_SZ_BYTES(fte_match_param)); dev_flow->ingress = attr->ingress; - dev_flow->transfer = attr->transfer; + dev_flow->dv.transfer = attr->transfer; return dev_flow; } -#ifndef NDEBUG +#ifdef RTE_LIBRTE_MLX5_DEBUG /** * Sanity check for match mask and value. Similar to check_valid_spec() in * kernel driver. If unmasked bit is present in value, it returns failure. @@ -5170,6 +5706,34 @@ flow_dv_check_valid_spec(void *match_mask, void *match_value) } #endif +/** + * Add match of ip_version. + * + * @param[in] group + * Flow group. + * @param[in] headers_v + * Values header pointer. + * @param[in] headers_m + * Masks header pointer. + * @param[in] ip_version + * The IP version to set. + */ +static inline void +flow_dv_set_match_ip_version(uint32_t group, + void *headers_v, + void *headers_m, + uint8_t ip_version) +{ + if (group == 0) + MLX5_SET(fte_match_set_lyr_2_4, headers_m, ip_version, 0xf); + else + MLX5_SET(fte_match_set_lyr_2_4, headers_m, ip_version, + ip_version); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_version, ip_version); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, ethertype, 0); + MLX5_SET(fte_match_set_lyr_2_4, headers_m, ethertype, 0); +} + /** * Add Ethernet item to matcher and to the value. * @@ -5184,7 +5748,8 @@ flow_dv_check_valid_spec(void *match_mask, void *match_value) */ static void flow_dv_translate_item_eth(void *matcher, void *key, - const struct rte_flow_item *item, int inner) + const struct rte_flow_item *item, int inner, + uint32_t group) { const struct rte_flow_item_eth *eth_m = item->mask; const struct rte_flow_item_eth *eth_v = item->spec; @@ -5223,10 +5788,38 @@ flow_dv_translate_item_eth(void *matcher, void *key, /* The value must be in the range of the mask. */ for (i = 0; i < sizeof(eth_m->dst); ++i) l24_v[i] = eth_m->src.addr_bytes[i] & eth_v->src.addr_bytes[i]; - MLX5_SET(fte_match_set_lyr_2_4, headers_m, ethertype, - rte_be_to_cpu_16(eth_m->type)); - l24_v = MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v, ethertype); - *(uint16_t *)(l24_v) = eth_m->type & eth_v->type; + if (eth_v->type) { + /* When ethertype is present set mask for tagged VLAN. */ + MLX5_SET(fte_match_set_lyr_2_4, headers_m, cvlan_tag, 1); + /* Set value for tagged VLAN if ethertype is 802.1Q. */ + if (eth_v->type == RTE_BE16(RTE_ETHER_TYPE_VLAN) || + eth_v->type == RTE_BE16(RTE_ETHER_TYPE_QINQ)) { + MLX5_SET(fte_match_set_lyr_2_4, headers_v, cvlan_tag, + 1); + /* Return here to avoid setting match on ethertype. */ + return; + } + } + /* + * HW supports match on one Ethertype, the Ethertype following the last + * VLAN tag of the packet (see PRM). + * Set match on ethertype only if ETH header is not followed by VLAN. + * HW is optimized for IPv4/IPv6. In such cases, avoid setting + * ethertype, and use ip_version field instead. + */ + if (eth_v->type == RTE_BE16(RTE_ETHER_TYPE_IPV4) && + eth_m->type == 0xFFFF) { + flow_dv_set_match_ip_version(group, headers_v, headers_m, 4); + } else if (eth_v->type == RTE_BE16(RTE_ETHER_TYPE_IPV6) && + eth_m->type == 0xFFFF) { + flow_dv_set_match_ip_version(group, headers_v, headers_m, 6); + } else { + MLX5_SET(fte_match_set_lyr_2_4, headers_m, ethertype, + rte_be_to_cpu_16(eth_m->type)); + l24_v = MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v, + ethertype); + *(uint16_t *)(l24_v) = eth_m->type & eth_v->type; + } } /** @@ -5247,7 +5840,7 @@ static void flow_dv_translate_item_vlan(struct mlx5_flow *dev_flow, void *matcher, void *key, const struct rte_flow_item *item, - int inner) + int inner, uint32_t group) { const struct rte_flow_item_vlan *vlan_m = item->mask; const struct rte_flow_item_vlan *vlan_v = item->spec; @@ -5256,10 +5849,6 @@ flow_dv_translate_item_vlan(struct mlx5_flow *dev_flow, uint16_t tci_m; uint16_t tci_v; - if (!vlan_v) - return; - if (!vlan_m) - vlan_m = &rte_flow_item_vlan_mask; if (inner) { headers_m = MLX5_ADDR_OF(fte_match_param, matcher, inner_headers); @@ -5272,23 +5861,45 @@ flow_dv_translate_item_vlan(struct mlx5_flow *dev_flow, * This is workaround, masks are not supported, * and pre-validated. */ - dev_flow->dv.vf_vlan.tag = - rte_be_to_cpu_16(vlan_v->tci) & 0x0fff; + if (vlan_v) + dev_flow->handle->vf_vlan.tag = + rte_be_to_cpu_16(vlan_v->tci) & 0x0fff; } - tci_m = rte_be_to_cpu_16(vlan_m->tci); - tci_v = rte_be_to_cpu_16(vlan_m->tci & vlan_v->tci); + /* + * When VLAN item exists in flow, mark packet as tagged, + * even if TCI is not specified. + */ MLX5_SET(fte_match_set_lyr_2_4, headers_m, cvlan_tag, 1); MLX5_SET(fte_match_set_lyr_2_4, headers_v, cvlan_tag, 1); + if (!vlan_v) + return; + if (!vlan_m) + vlan_m = &rte_flow_item_vlan_mask; + tci_m = rte_be_to_cpu_16(vlan_m->tci); + tci_v = rte_be_to_cpu_16(vlan_m->tci & vlan_v->tci); MLX5_SET(fte_match_set_lyr_2_4, headers_m, first_vid, tci_m); MLX5_SET(fte_match_set_lyr_2_4, headers_v, first_vid, tci_v); MLX5_SET(fte_match_set_lyr_2_4, headers_m, first_cfi, tci_m >> 12); MLX5_SET(fte_match_set_lyr_2_4, headers_v, first_cfi, tci_v >> 12); MLX5_SET(fte_match_set_lyr_2_4, headers_m, first_prio, tci_m >> 13); MLX5_SET(fte_match_set_lyr_2_4, headers_v, first_prio, tci_v >> 13); - MLX5_SET(fte_match_set_lyr_2_4, headers_m, ethertype, - rte_be_to_cpu_16(vlan_m->inner_type)); - MLX5_SET(fte_match_set_lyr_2_4, headers_v, ethertype, - rte_be_to_cpu_16(vlan_m->inner_type & vlan_v->inner_type)); + /* + * HW is optimized for IPv4/IPv6. In such cases, avoid setting + * ethertype, and use ip_version field instead. + */ + if (vlan_v->inner_type == RTE_BE16(RTE_ETHER_TYPE_IPV4) && + vlan_m->inner_type == 0xFFFF) { + flow_dv_set_match_ip_version(group, headers_v, headers_m, 4); + } else if (vlan_v->inner_type == RTE_BE16(RTE_ETHER_TYPE_IPV6) && + vlan_m->inner_type == 0xFFFF) { + flow_dv_set_match_ip_version(group, headers_v, headers_m, 6); + } else { + MLX5_SET(fte_match_set_lyr_2_4, headers_m, ethertype, + rte_be_to_cpu_16(vlan_m->inner_type)); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, ethertype, + rte_be_to_cpu_16(vlan_m->inner_type & + vlan_v->inner_type)); + } } /** @@ -5300,6 +5911,8 @@ flow_dv_translate_item_vlan(struct mlx5_flow *dev_flow, * Flow matcher value. * @param[in] item * Flow pattern to translate. + * @param[in] item_flags + * Bit-fields that holds the items detected until now. * @param[in] inner * Item is inner pattern. * @param[in] group @@ -5308,6 +5921,7 @@ flow_dv_translate_item_vlan(struct mlx5_flow *dev_flow, static void flow_dv_translate_item_ipv4(void *matcher, void *key, const struct rte_flow_item *item, + const uint64_t item_flags, int inner, uint32_t group) { const struct rte_flow_item_ipv4 *ipv4_m = item->mask; @@ -5318,6 +5932,7 @@ flow_dv_translate_item_ipv4(void *matcher, void *key, .dst_addr = RTE_BE32(0xffffffff), .type_of_service = 0xff, .next_proto_id = 0xff, + .time_to_live = 0xff, }, }; void *headers_m; @@ -5335,11 +5950,14 @@ flow_dv_translate_item_ipv4(void *matcher, void *key, outer_headers); headers_v = MLX5_ADDR_OF(fte_match_param, key, outer_headers); } - if (group == 0) - MLX5_SET(fte_match_set_lyr_2_4, headers_m, ip_version, 0xf); - else - MLX5_SET(fte_match_set_lyr_2_4, headers_m, ip_version, 0x4); - MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_version, 4); + flow_dv_set_match_ip_version(group, headers_v, headers_m, 4); + /* + * On outer header (which must contains L2), or inner header with L2, + * set cvlan_tag mask bit to mark this packet as untagged. + * This should be done even if item->spec is empty. + */ + if (!inner || item_flags & MLX5_FLOW_LAYER_INNER_L2) + MLX5_SET(fte_match_set_lyr_2_4, headers_m, cvlan_tag, 1); if (!ipv4_v) return; if (!ipv4_m) @@ -5367,6 +5985,10 @@ flow_dv_translate_item_ipv4(void *matcher, void *key, ipv4_m->hdr.next_proto_id); MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_protocol, ipv4_v->hdr.next_proto_id & ipv4_m->hdr.next_proto_id); + MLX5_SET(fte_match_set_lyr_2_4, headers_m, ip_ttl_hoplimit, + ipv4_m->hdr.time_to_live); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_ttl_hoplimit, + ipv4_v->hdr.time_to_live & ipv4_m->hdr.time_to_live); } /** @@ -5378,6 +6000,8 @@ flow_dv_translate_item_ipv4(void *matcher, void *key, * Flow matcher value. * @param[in] item * Flow pattern to translate. + * @param[in] item_flags + * Bit-fields that holds the items detected until now. * @param[in] inner * Item is inner pattern. * @param[in] group @@ -5386,6 +6010,7 @@ flow_dv_translate_item_ipv4(void *matcher, void *key, static void flow_dv_translate_item_ipv6(void *matcher, void *key, const struct rte_flow_item *item, + const uint64_t item_flags, int inner, uint32_t group) { const struct rte_flow_item_ipv6 *ipv6_m = item->mask; @@ -5423,11 +6048,14 @@ flow_dv_translate_item_ipv6(void *matcher, void *key, outer_headers); headers_v = MLX5_ADDR_OF(fte_match_param, key, outer_headers); } - if (group == 0) - MLX5_SET(fte_match_set_lyr_2_4, headers_m, ip_version, 0xf); - else - MLX5_SET(fte_match_set_lyr_2_4, headers_m, ip_version, 0x6); - MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_version, 6); + flow_dv_set_match_ip_version(group, headers_v, headers_m, 6); + /* + * On outer header (which must contains L2), or inner header with L2, + * set cvlan_tag mask bit to mark this packet as untagged. + * This should be done even if item->spec is empty. + */ + if (!inner || item_flags & MLX5_FLOW_LAYER_INNER_L2) + MLX5_SET(fte_match_set_lyr_2_4, headers_m, cvlan_tag, 1); if (!ipv6_v) return; if (!ipv6_m) @@ -5471,6 +6099,11 @@ flow_dv_translate_item_ipv6(void *matcher, void *key, ipv6_m->hdr.proto); MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_protocol, ipv6_v->hdr.proto & ipv6_m->hdr.proto); + /* Hop limit. */ + MLX5_SET(fte_match_set_lyr_2_4, headers_m, ip_ttl_hoplimit, + ipv6_m->hdr.hop_limits); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_ttl_hoplimit, + ipv6_v->hdr.hop_limits & ipv6_m->hdr.hop_limits); } /** @@ -5593,13 +6226,13 @@ flow_dv_translate_item_gre_key(void *matcher, void *key, void *misc_v = MLX5_ADDR_OF(fte_match_param, key, misc_parameters); rte_be32_t gre_key_default_mask = RTE_BE32(UINT32_MAX); + /* GRE K bit must be on and should already be validated */ + MLX5_SET(fte_match_set_misc, misc_m, gre_k_present, 1); + MLX5_SET(fte_match_set_misc, misc_v, gre_k_present, 1); if (!key_v) return; if (!key_m) key_m = &gre_key_default_mask; - /* GRE K bit must be on and should already be validated */ - MLX5_SET(fte_match_set_misc, misc_m, gre_k_present, 1); - MLX5_SET(fte_match_set_misc, misc_v, gre_k_present, 1); MLX5_SET(fte_match_set_misc, misc_m, gre_key_h, rte_be_to_cpu_32(*key_m) >> 8); MLX5_SET(fte_match_set_misc, misc_v, gre_key_h, @@ -5797,6 +6430,76 @@ flow_dv_translate_item_vxlan(void *matcher, void *key, vni_v[i] = vni_m[i] & vxlan_v->vni[i]; } +/** + * Add VXLAN-GPE item to matcher and to the value. + * + * @param[in, out] matcher + * Flow matcher. + * @param[in, out] key + * Flow matcher value. + * @param[in] item + * Flow pattern to translate. + * @param[in] inner + * Item is inner pattern. + */ + +static void +flow_dv_translate_item_vxlan_gpe(void *matcher, void *key, + const struct rte_flow_item *item, int inner) +{ + const struct rte_flow_item_vxlan_gpe *vxlan_m = item->mask; + const struct rte_flow_item_vxlan_gpe *vxlan_v = item->spec; + void *headers_m; + void *headers_v; + void *misc_m = + MLX5_ADDR_OF(fte_match_param, matcher, misc_parameters_3); + void *misc_v = + MLX5_ADDR_OF(fte_match_param, key, misc_parameters_3); + char *vni_m; + char *vni_v; + uint16_t dport; + int size; + int i; + uint8_t flags_m = 0xff; + uint8_t flags_v = 0xc; + + if (inner) { + headers_m = MLX5_ADDR_OF(fte_match_param, matcher, + inner_headers); + headers_v = MLX5_ADDR_OF(fte_match_param, key, inner_headers); + } else { + headers_m = MLX5_ADDR_OF(fte_match_param, matcher, + outer_headers); + headers_v = MLX5_ADDR_OF(fte_match_param, key, outer_headers); + } + dport = item->type == RTE_FLOW_ITEM_TYPE_VXLAN ? + MLX5_UDP_PORT_VXLAN : MLX5_UDP_PORT_VXLAN_GPE; + if (!MLX5_GET16(fte_match_set_lyr_2_4, headers_v, udp_dport)) { + MLX5_SET(fte_match_set_lyr_2_4, headers_m, udp_dport, 0xFFFF); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, udp_dport, dport); + } + if (!vxlan_v) + return; + if (!vxlan_m) + vxlan_m = &rte_flow_item_vxlan_gpe_mask; + size = sizeof(vxlan_m->vni); + vni_m = MLX5_ADDR_OF(fte_match_set_misc3, misc_m, outer_vxlan_gpe_vni); + vni_v = MLX5_ADDR_OF(fte_match_set_misc3, misc_v, outer_vxlan_gpe_vni); + memcpy(vni_m, vxlan_m->vni, size); + for (i = 0; i < size; ++i) + vni_v[i] = vni_m[i] & vxlan_v->vni[i]; + if (vxlan_m->flags) { + flags_m = vxlan_m->flags; + flags_v = vxlan_v->flags; + } + MLX5_SET(fte_match_set_misc3, misc_m, outer_vxlan_gpe_flags, flags_m); + MLX5_SET(fte_match_set_misc3, misc_v, outer_vxlan_gpe_flags, flags_v); + MLX5_SET(fte_match_set_misc3, misc_m, outer_vxlan_gpe_next_protocol, + vxlan_m->protocol); + MLX5_SET(fte_match_set_misc3, misc_v, outer_vxlan_gpe_next_protocol, + vxlan_v->protocol); +} + /** * Add Geneve item to matcher and to the value. * @@ -6036,7 +6739,7 @@ flow_dv_match_meta_reg(void *matcher, void *key, MLX5_SET(fte_match_set_misc2, misc2_v, metadata_reg_c_7, data); break; default: - assert(false); + MLX5_ASSERT(false); break; } } @@ -6067,14 +6770,14 @@ flow_dv_translate_item_mark(struct rte_eth_dev *dev, &rte_flow_item_mark_mask; mask = mark->id & priv->sh->dv_mark_mask; mark = (const void *)item->spec; - assert(mark); + MLX5_ASSERT(mark); value = mark->id & priv->sh->dv_mark_mask & mask; if (mask) { enum modify_reg reg; /* Get the metadata register index for the mark. */ reg = mlx5_flow_get_reg_id(dev, MLX5_FLOW_MARK, 0, NULL); - assert(reg > 0); + MLX5_ASSERT(reg > 0); if (reg == REG_C_0) { struct mlx5_priv *priv = dev->data->dev_private; uint32_t msk_c0 = priv->sh->dv_regc0_mask; @@ -6116,7 +6819,7 @@ flow_dv_translate_item_meta(struct rte_eth_dev *dev, meta_m = &rte_flow_item_meta_mask; meta_v = (const void *)item->spec; if (meta_v) { - enum modify_reg reg; + int reg; uint32_t value = meta_v->data; uint32_t mask = meta_m->data; @@ -6142,8 +6845,8 @@ flow_dv_translate_item_meta(struct rte_eth_dev *dev, #endif value <<= shl_c0; mask <<= shl_c0; - assert(msk_c0); - assert(!(~msk_c0 & mask)); + MLX5_ASSERT(msk_c0); + MLX5_ASSERT(!(~msk_c0 & mask)); } flow_dv_match_meta_reg(matcher, key, reg, value, mask); } @@ -6187,7 +6890,7 @@ flow_dv_translate_mlx5_item_tag(struct rte_eth_dev *dev, const struct mlx5_rte_flow_item_tag *tag_m = item->mask; uint32_t mask, value; - assert(tag_v); + MLX5_ASSERT(tag_v); value = tag_v->data; mask = tag_m ? tag_m->data : UINT32_MAX; if (tag_v->id == REG_C_0) { @@ -6223,11 +6926,11 @@ flow_dv_translate_item_tag(struct rte_eth_dev *dev, const struct rte_flow_item_tag *tag_m = item->mask; enum modify_reg reg; - assert(tag_v); + MLX5_ASSERT(tag_v); tag_m = tag_m ? tag_m : &rte_flow_item_tag_mask; /* Get the metadata register index for the tag. */ reg = mlx5_flow_get_reg_id(dev, MLX5_APP_TAG, tag_v->index, NULL); - assert(reg > 0); + MLX5_ASSERT(reg > 0); flow_dv_match_meta_reg(matcher, key, reg, tag_v->data, tag_m->data); } @@ -6333,6 +7036,12 @@ flow_dv_translate_item_icmp6(void *matcher, void *key, return; if (!icmp6_m) icmp6_m = &rte_flow_item_icmp6_mask; + /* + * Force flow only to match the non-fragmented IPv6 ICMPv6 packets. + * If only the protocol is specified, no need to match the frag. + */ + MLX5_SET(fte_match_set_lyr_2_4, headers_m, frag, 1); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, frag, 0); MLX5_SET(fte_match_set_misc3, misc3_m, icmpv6_type, icmp6_m->type); MLX5_SET(fte_match_set_misc3, misc3_v, icmpv6_type, icmp6_v->type & icmp6_m->type); @@ -6380,6 +7089,12 @@ flow_dv_translate_item_icmp(void *matcher, void *key, return; if (!icmp_m) icmp_m = &rte_flow_item_icmp_mask; + /* + * Force flow only to match the non-fragmented IPv4 ICMP packets. + * If only the protocol is specified, no need to match the frag. + */ + MLX5_SET(fte_match_set_lyr_2_4, headers_m, frag, 1); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, frag, 0); MLX5_SET(fte_match_set_misc3, misc3_m, icmp_type, icmp_m->hdr.icmp_type); MLX5_SET(fte_match_set_misc3, misc3_v, icmp_type, @@ -6432,6 +7147,10 @@ flow_dv_translate_item_gtp(void *matcher, void *key, return; if (!gtp_m) gtp_m = &rte_flow_item_gtp_mask; + MLX5_SET(fte_match_set_misc3, misc3_m, gtpu_msg_flags, + gtp_m->v_pt_rsv_flags); + MLX5_SET(fte_match_set_misc3, misc3_v, gtpu_msg_flags, + gtp_v->v_pt_rsv_flags & gtp_m->v_pt_rsv_flags); MLX5_SET(fte_match_set_misc3, misc3_m, gtpu_msg_type, gtp_m->msg_type); MLX5_SET(fte_match_set_misc3, misc3_v, gtpu_msg_type, gtp_v->msg_type & gtp_m->msg_type); @@ -6517,6 +7236,7 @@ flow_dv_tbl_resource_get(struct rte_eth_dev *dev, struct mlx5_hlist_entry *pos = mlx5_hlist_lookup(sh->flow_tbls, table_key.v64); struct mlx5_flow_tbl_data_entry *tbl_data; + uint32_t idx = 0; int ret; void *domain; @@ -6527,7 +7247,7 @@ flow_dv_tbl_resource_get(struct rte_eth_dev *dev, rte_atomic32_inc(&tbl->refcnt); return tbl; } - tbl_data = rte_zmalloc(NULL, sizeof(*tbl_data), 0); + tbl_data = mlx5_ipool_zmalloc(sh->ipool[MLX5_IPOOL_JUMP], &idx); if (!tbl_data) { rte_flow_error_set(error, ENOMEM, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, @@ -6535,6 +7255,7 @@ flow_dv_tbl_resource_get(struct rte_eth_dev *dev, "cannot allocate flow table data entry"); return NULL; } + tbl_data->idx = idx; tbl = &tbl_data->tbl; pos = &tbl_data->entry; if (transfer) @@ -6548,7 +7269,7 @@ flow_dv_tbl_resource_get(struct rte_eth_dev *dev, rte_flow_error_set(error, ENOMEM, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, "cannot create flow table object"); - rte_free(tbl_data); + mlx5_ipool_free(sh->ipool[MLX5_IPOOL_JUMP], idx); return NULL; } /* @@ -6565,7 +7286,7 @@ flow_dv_tbl_resource_get(struct rte_eth_dev *dev, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, "cannot insert flow table data entry"); mlx5_glue->dr_destroy_flow_tbl(tbl->obj); - rte_free(tbl_data); + mlx5_ipool_free(sh->ipool[MLX5_IPOOL_JUMP], idx); } rte_atomic32_inc(&tbl->refcnt); return tbl; @@ -6600,7 +7321,8 @@ flow_dv_tbl_resource_release(struct rte_eth_dev *dev, tbl->obj = NULL; /* remove the entry from the hash list and free memory. */ mlx5_hlist_remove(sh->flow_tbls, pos); - rte_free(tbl_data); + mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_JUMP], + tbl_data->idx); return 0; } return 1; @@ -6661,7 +7383,7 @@ flow_dv_matcher_register(struct rte_eth_dev *dev, (void *)cache_matcher, rte_atomic32_read(&cache_matcher->refcnt)); rte_atomic32_inc(&cache_matcher->refcnt); - dev_flow->dv.matcher = cache_matcher; + dev_flow->handle->dvh.matcher = cache_matcher; /* old matcher should not make the table ref++. */ flow_dv_tbl_resource_release(dev, tbl); return 0; @@ -6698,7 +7420,7 @@ flow_dv_matcher_register(struct rte_eth_dev *dev, /* only matcher ref++, table ref++ already done above in get API. */ rte_atomic32_inc(&cache_matcher->refcnt); LIST_INSERT_HEAD(&tbl_data->matchers, cache_matcher, next); - dev_flow->dv.matcher = cache_matcher; + dev_flow->handle->dvh.matcher = cache_matcher; DRV_LOG(DEBUG, "%s group %u priority %hd new %s matcher %p: refcnt %d", key->domain ? "FDB" : "NIC", key->table_id, cache_matcher->priority, @@ -6740,6 +7462,7 @@ flow_dv_tag_resource_register cache_resource = container_of (entry, struct mlx5_flow_dv_tag_resource, entry); rte_atomic32_inc(&cache_resource->refcnt); + dev_flow->handle->dvh.rix_tag = cache_resource->idx; dev_flow->dv.tag_resource = cache_resource; DRV_LOG(DEBUG, "cached tag resource %p: refcnt now %d++", (void *)cache_resource, @@ -6747,7 +7470,8 @@ flow_dv_tag_resource_register return 0; } /* Register new resource. */ - cache_resource = rte_calloc(__func__, 1, sizeof(*cache_resource), 0); + cache_resource = mlx5_ipool_zmalloc(sh->ipool[MLX5_IPOOL_TAG], + &dev_flow->handle->dvh.rix_tag); if (!cache_resource) return rte_flow_error_set(error, ENOMEM, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, @@ -6781,20 +7505,23 @@ flow_dv_tag_resource_register * * @param dev * Pointer to Ethernet device. - * @param flow - * Pointer to mlx5_flow. + * @param tag_idx + * Tag index. * * @return * 1 while a reference on it exists, 0 when freed. */ static int flow_dv_tag_release(struct rte_eth_dev *dev, - struct mlx5_flow_dv_tag_resource *tag) + uint32_t tag_idx) { struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_ibv_shared *sh = priv->sh; + struct mlx5_flow_dv_tag_resource *tag; - assert(tag); + tag = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_TAG], tag_idx); + if (!tag) + return 0; DRV_LOG(DEBUG, "port %u tag %p: refcnt %d--", dev->data->port_id, (void *)tag, rte_atomic32_read(&tag->refcnt)); @@ -6803,7 +7530,7 @@ flow_dv_tag_release(struct rte_eth_dev *dev, mlx5_hlist_remove(sh->tag_table, &tag->entry); DRV_LOG(DEBUG, "port %u tag %p: removed", dev->data->port_id, (void *)tag); - rte_free(tag); + mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_TAG], tag_idx); return 0; } return 1; @@ -6859,6 +7586,53 @@ flow_dv_translate_action_port_id(struct rte_eth_dev *dev, return 0; } +/** + * Create a counter with aging configuration. + * + * @param[in] dev + * Pointer to rte_eth_dev structure. + * @param[out] count + * Pointer to the counter action configuration. + * @param[in] age + * Pointer to the aging action configuration. + * + * @return + * Index to flow counter on success, 0 otherwise. + */ +static uint32_t +flow_dv_translate_create_counter(struct rte_eth_dev *dev, + struct mlx5_flow *dev_flow, + const struct rte_flow_action_count *count, + const struct rte_flow_action_age *age) +{ + uint32_t counter; + struct mlx5_age_param *age_param; + + counter = flow_dv_counter_alloc(dev, + count ? count->shared : 0, + count ? count->id : 0, + dev_flow->dv.group, !!age); + if (!counter || age == NULL) + return counter; + age_param = flow_dv_counter_idx_get_age(dev, counter); + /* + * The counter age accuracy may have a bit delay. Have 3/4 + * second bias on the timeount in order to let it age in time. + */ + age_param->context = age->context ? age->context : + (void *)(uintptr_t)(dev_flow->flow_idx); + /* + * The counter age accuracy may have a bit delay. Have 3/4 + * second bias on the timeount in order to let it age in time. + */ + age_param->timeout = age->timeout * 10 - MLX5_AGING_TIME_DELAY; + /* Set expire time in unit of 0.1 sec. */ + age_param->port_id = dev->data->port_id; + age_param->expire = age_param->timeout + + rte_rdtsc() / (rte_get_tsc_hz() / 10); + rte_atomic16_set(&age_param->state, AGE_CANDIDATE); + return counter; +} /** * Add Tx queue matcher * @@ -6909,18 +7683,20 @@ flow_dv_translate_item_tx_queue(struct rte_eth_dev *dev, * * @param[in] dev_flow * Pointer to the mlx5_flow. + * @param[in] rss_desc + * Pointer to the mlx5_flow_rss_desc. */ static void -flow_dv_hashfields_set(struct mlx5_flow *dev_flow) +flow_dv_hashfields_set(struct mlx5_flow *dev_flow, + struct mlx5_flow_rss_desc *rss_desc) { - struct rte_flow *flow = dev_flow->flow; - uint64_t items = dev_flow->layers; + uint64_t items = dev_flow->handle->layers; int rss_inner = 0; - uint64_t rss_types = rte_eth_rss_hf_refine(flow->rss.types); + uint64_t rss_types = rte_eth_rss_hf_refine(rss_desc->types); dev_flow->hash_fields = 0; #ifdef HAVE_IBV_DEVICE_TUNNEL_SUPPORT - if (flow->rss.level >= 2) { + if (rss_desc->level >= 2) { dev_flow->hash_fields |= IBV_RX_HASH_INNER; rss_inner = 1; } @@ -7004,6 +7780,10 @@ __flow_dv_translate(struct rte_eth_dev *dev, struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_dev_config *dev_conf = &priv->config; struct rte_flow *flow = dev_flow->flow; + struct mlx5_flow_handle *handle = dev_flow->handle; + struct mlx5_flow_rss_desc *rss_desc = &((struct mlx5_flow_rss_desc *) + priv->rss_desc) + [!!priv->flow_nested_idx]; uint64_t item_flags = 0; uint64_t last_item = 0; uint64_t action_flags = 0; @@ -7022,6 +7802,8 @@ __flow_dv_translate(struct rte_eth_dev *dev, (MLX5_MAX_MODIFY_NUM + 1)]; } mhdr_dummy; struct mlx5_flow_dv_modify_hdr_resource *mhdr_res = &mhdr_dummy.res; + const struct rte_flow_action_count *count = NULL; + const struct rte_flow_action_age *age = NULL; union flow_dv_attr flow_attr = { .attr = 0 }; uint32_t tag_be; union mlx5_flow_tbl_key tbl_key; @@ -7036,10 +7818,10 @@ __flow_dv_translate(struct rte_eth_dev *dev, mhdr_res->ft_type = attr->egress ? MLX5DV_FLOW_TABLE_TYPE_NIC_TX : MLX5DV_FLOW_TABLE_TYPE_NIC_RX; ret = mlx5_flow_group_to_table(attr, dev_flow->external, attr->group, - &table, error); + !!priv->fdb_def_rule, &table, error); if (ret) return ret; - dev_flow->group = table; + dev_flow->dv.group = table; if (attr->transfer) mhdr_res->ft_type = MLX5DV_FLOW_TABLE_TYPE_FDB; if (priority == MLX5_FLOW_PRIO_RSVD) @@ -7050,7 +7832,6 @@ __flow_dv_translate(struct rte_eth_dev *dev, const struct rte_flow_action_queue *queue; const struct rte_flow_action_rss *rss; const struct rte_flow_action *action = actions; - const struct rte_flow_action_count *count = action->conf; const uint8_t *rss_key; const struct rte_flow_action_jump *jump_data; const struct rte_flow_action_meter *mtr; @@ -7059,6 +7840,7 @@ __flow_dv_translate(struct rte_eth_dev *dev, struct mlx5_flow_dv_port_id_action_resource port_id_resource; int action_type = actions->type; const struct rte_flow_action *found_action = NULL; + struct mlx5_flow_meter *fm = NULL; switch (action_type) { case RTE_FLOW_ACTION_TYPE_VOID: @@ -7067,16 +7849,20 @@ __flow_dv_translate(struct rte_eth_dev *dev, if (flow_dv_translate_action_port_id(dev, action, &port_id, error)) return -rte_errno; + memset(&port_id_resource, 0, sizeof(port_id_resource)); port_id_resource.port_id = port_id; if (flow_dv_port_id_action_resource_register (dev, &port_id_resource, dev_flow, error)) return -rte_errno; + MLX5_ASSERT(!handle->rix_port_id_action); dev_flow->dv.actions[actions_n++] = - dev_flow->dv.port_id_action->action; + dev_flow->dv.port_id_action->action; action_flags |= MLX5_FLOW_ACTION_PORT_ID; + dev_flow->handle->fate_action = MLX5_FLOW_FATE_PORT_ID; break; case RTE_FLOW_ACTION_TYPE_FLAG: action_flags |= MLX5_FLOW_ACTION_FLAG; + dev_flow->handle->mark = 1; if (dev_conf->dv_xmeta_en != MLX5_XMETA_MODE_LEGACY) { struct rte_flow_action_mark mark = { .id = MLX5_FLOW_MARK_DEFAULT, @@ -7090,15 +7876,22 @@ __flow_dv_translate(struct rte_eth_dev *dev, break; } tag_be = mlx5_flow_mark_set(MLX5_FLOW_MARK_DEFAULT); - if (!dev_flow->dv.tag_resource) - if (flow_dv_tag_resource_register - (dev, tag_be, dev_flow, error)) - return -rte_errno; + /* + * Only one FLAG or MARK is supported per device flow + * right now. So the pointer to the tag resource must be + * zero before the register process. + */ + MLX5_ASSERT(!handle->dvh.rix_tag); + if (flow_dv_tag_resource_register(dev, tag_be, + dev_flow, error)) + return -rte_errno; + MLX5_ASSERT(dev_flow->dv.tag_resource); dev_flow->dv.actions[actions_n++] = - dev_flow->dv.tag_resource->action; + dev_flow->dv.tag_resource->action; break; case RTE_FLOW_ACTION_TYPE_MARK: action_flags |= MLX5_FLOW_ACTION_MARK; + dev_flow->handle->mark = 1; if (dev_conf->dv_xmeta_en != MLX5_XMETA_MODE_LEGACY) { const struct rte_flow_action_mark *mark = (const struct rte_flow_action_mark *) @@ -7117,12 +7910,13 @@ __flow_dv_translate(struct rte_eth_dev *dev, tag_be = mlx5_flow_mark_set (((const struct rte_flow_action_mark *) (actions->conf))->id); - if (!dev_flow->dv.tag_resource) - if (flow_dv_tag_resource_register - (dev, tag_be, dev_flow, error)) - return -rte_errno; + MLX5_ASSERT(!handle->dvh.rix_tag); + if (flow_dv_tag_resource_register(dev, tag_be, + dev_flow, error)) + return -rte_errno; + MLX5_ASSERT(dev_flow->dv.tag_resource); dev_flow->dv.actions[actions_n++] = - dev_flow->dv.tag_resource->action; + dev_flow->dv.tag_resource->action; break; case RTE_FLOW_ACTION_TYPE_SET_META: if (flow_dv_convert_action_set_meta @@ -7142,59 +7936,45 @@ __flow_dv_translate(struct rte_eth_dev *dev, break; case RTE_FLOW_ACTION_TYPE_DROP: action_flags |= MLX5_FLOW_ACTION_DROP; + dev_flow->handle->fate_action = MLX5_FLOW_FATE_DROP; break; case RTE_FLOW_ACTION_TYPE_QUEUE: - assert(flow->rss.queue); queue = actions->conf; - flow->rss.queue_num = 1; - (*flow->rss.queue)[0] = queue->index; + rss_desc->queue_num = 1; + rss_desc->queue[0] = queue->index; action_flags |= MLX5_FLOW_ACTION_QUEUE; + dev_flow->handle->fate_action = MLX5_FLOW_FATE_QUEUE; break; case RTE_FLOW_ACTION_TYPE_RSS: - assert(flow->rss.queue); rss = actions->conf; - if (flow->rss.queue) - memcpy((*flow->rss.queue), rss->queue, - rss->queue_num * sizeof(uint16_t)); - flow->rss.queue_num = rss->queue_num; + 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(flow->rss.key, rss_key, MLX5_RSS_HASH_KEY_LEN); + 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. */ action_flags |= MLX5_FLOW_ACTION_RSS; + dev_flow->handle->fate_action = MLX5_FLOW_FATE_QUEUE; break; + case RTE_FLOW_ACTION_TYPE_AGE: case RTE_FLOW_ACTION_TYPE_COUNT: if (!dev_conf->devx) { - rte_errno = ENOTSUP; - goto cnt_err; - } - flow->counter = flow_dv_counter_alloc(dev, - count->shared, - count->id, - dev_flow->group); - if (flow->counter == NULL) - goto cnt_err; - dev_flow->dv.actions[actions_n++] = - flow->counter->action; - action_flags |= MLX5_FLOW_ACTION_COUNT; - break; -cnt_err: - if (rte_errno == ENOTSUP) return rte_flow_error_set (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, "count action not supported"); + } + /* Save information first, will apply later. */ + if (actions->type == RTE_FLOW_ACTION_TYPE_COUNT) + count = action->conf; else - return rte_flow_error_set - (error, rte_errno, - RTE_FLOW_ERROR_TYPE_ACTION, - action, - "cannot create counter" - " object."); + age = action->conf; + action_flags |= MLX5_FLOW_ACTION_COUNT; break; case RTE_FLOW_ACTION_TYPE_OF_POP_VLAN: dev_flow->dv.actions[actions_n++] = @@ -7202,7 +7982,9 @@ cnt_err: action_flags |= MLX5_FLOW_ACTION_OF_POP_VLAN; break; case RTE_FLOW_ACTION_TYPE_OF_PUSH_VLAN: - flow_dev_get_vlan_info_from_items(items, &vlan); + if (!(action_flags & + MLX5_FLOW_ACTION_OF_SET_VLAN_VID)) + flow_dev_get_vlan_info_from_items(items, &vlan); vlan.eth_proto = rte_be_to_cpu_16 ((((const struct rte_flow_action_of_push_vlan *) actions->conf)->ethertype)); @@ -7220,12 +8002,13 @@ cnt_err: (dev, attr, &vlan, dev_flow, error)) return -rte_errno; dev_flow->dv.actions[actions_n++] = - dev_flow->dv.push_vlan_res->action; + dev_flow->dv.push_vlan_res->action; action_flags |= MLX5_FLOW_ACTION_OF_PUSH_VLAN; break; case RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_PCP: /* of_vlan_push action handled this action */ - assert(action_flags & MLX5_FLOW_ACTION_OF_PUSH_VLAN); + MLX5_ASSERT(action_flags & + MLX5_FLOW_ACTION_OF_PUSH_VLAN); break; case RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_VID: if (action_flags & MLX5_FLOW_ACTION_OF_PUSH_VLAN) @@ -7246,11 +8029,8 @@ cnt_err: error)) return -rte_errno; dev_flow->dv.actions[actions_n++] = - dev_flow->dv.encap_decap->verbs_action; - action_flags |= actions->type == - RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP ? - MLX5_FLOW_ACTION_VXLAN_ENCAP : - MLX5_FLOW_ACTION_NVGRE_ENCAP; + dev_flow->dv.encap_decap->verbs_action; + action_flags |= MLX5_FLOW_ACTION_ENCAP; break; case RTE_FLOW_ACTION_TYPE_VXLAN_DECAP: case RTE_FLOW_ACTION_TYPE_NVGRE_DECAP: @@ -7259,15 +8039,12 @@ cnt_err: error)) return -rte_errno; dev_flow->dv.actions[actions_n++] = - dev_flow->dv.encap_decap->verbs_action; - action_flags |= actions->type == - RTE_FLOW_ACTION_TYPE_VXLAN_DECAP ? - MLX5_FLOW_ACTION_VXLAN_DECAP : - MLX5_FLOW_ACTION_NVGRE_DECAP; + dev_flow->dv.encap_decap->verbs_action; + action_flags |= MLX5_FLOW_ACTION_DECAP; break; case RTE_FLOW_ACTION_TYPE_RAW_ENCAP: /* Handle encap with preceding decap. */ - if (action_flags & MLX5_FLOW_ACTION_RAW_DECAP) { + if (action_flags & MLX5_FLOW_ACTION_DECAP) { if (flow_dv_create_action_raw_encap (dev, actions, dev_flow, attr, error)) return -rte_errno; @@ -7282,15 +8059,11 @@ cnt_err: dev_flow->dv.actions[actions_n++] = dev_flow->dv.encap_decap->verbs_action; } - action_flags |= MLX5_FLOW_ACTION_RAW_ENCAP; + action_flags |= MLX5_FLOW_ACTION_ENCAP; break; case RTE_FLOW_ACTION_TYPE_RAW_DECAP: - /* Check if this decap is followed by encap. */ - for (; action->type != RTE_FLOW_ACTION_TYPE_END && - action->type != RTE_FLOW_ACTION_TYPE_RAW_ENCAP; - action++) { - } - /* Handle decap only if it isn't followed by encap. */ + while ((++action)->type == RTE_FLOW_ACTION_TYPE_VOID) + ; if (action->type != RTE_FLOW_ACTION_TYPE_RAW_ENCAP) { if (flow_dv_create_action_l2_decap (dev, dev_flow, attr->transfer, error)) @@ -7299,13 +8072,14 @@ cnt_err: dev_flow->dv.encap_decap->verbs_action; } /* If decap is followed by encap, handle it at encap. */ - action_flags |= MLX5_FLOW_ACTION_RAW_DECAP; + action_flags |= MLX5_FLOW_ACTION_DECAP; break; case RTE_FLOW_ACTION_TYPE_JUMP: jump_data = action->conf; ret = mlx5_flow_group_to_table(attr, dev_flow->external, - jump_data->group, &table, - error); + jump_data->group, + !!priv->fdb_def_rule, + &table, error); if (ret) return ret; tbl = flow_dv_tbl_resource_get(dev, table, @@ -7327,8 +8101,9 @@ cnt_err: "cannot create jump action."); } dev_flow->dv.actions[actions_n++] = - dev_flow->dv.jump->action; + dev_flow->dv.jump->action; action_flags |= MLX5_FLOW_ACTION_JUMP; + dev_flow->handle->fate_action = MLX5_FLOW_FATE_JUMP; break; case RTE_FLOW_ACTION_TYPE_SET_MAC_SRC: case RTE_FLOW_ACTION_TYPE_SET_MAC_DST: @@ -7364,7 +8139,8 @@ cnt_err: case RTE_FLOW_ACTION_TYPE_SET_TP_DST: if (flow_dv_convert_action_modify_tp (mhdr_res, actions, items, - &flow_attr, error)) + &flow_attr, dev_flow, !!(action_flags & + MLX5_FLOW_ACTION_DECAP), error)) return -rte_errno; action_flags |= actions->type == RTE_FLOW_ACTION_TYPE_SET_TP_SRC ? @@ -7373,14 +8149,17 @@ cnt_err: break; case RTE_FLOW_ACTION_TYPE_DEC_TTL: if (flow_dv_convert_action_modify_dec_ttl - (mhdr_res, items, &flow_attr, error)) + (mhdr_res, items, &flow_attr, dev_flow, + !!(action_flags & + MLX5_FLOW_ACTION_DECAP), error)) return -rte_errno; action_flags |= MLX5_FLOW_ACTION_DEC_TTL; break; case RTE_FLOW_ACTION_TYPE_SET_TTL: if (flow_dv_convert_action_modify_ttl - (mhdr_res, actions, items, - &flow_attr, error)) + (mhdr_res, actions, items, &flow_attr, + dev_flow, !!(action_flags & + MLX5_FLOW_ACTION_DECAP), error)) return -rte_errno; action_flags |= MLX5_FLOW_ACTION_SET_TTL; break; @@ -7420,20 +8199,31 @@ cnt_err: case RTE_FLOW_ACTION_TYPE_METER: mtr = actions->conf; if (!flow->meter) { - flow->meter = mlx5_flow_meter_attach(priv, - mtr->mtr_id, attr, - error); - if (!flow->meter) + fm = mlx5_flow_meter_attach(priv, mtr->mtr_id, + attr, error); + if (!fm) return rte_flow_error_set(error, rte_errno, RTE_FLOW_ERROR_TYPE_ACTION, NULL, "meter not found " "or invalid parameters"); + flow->meter = fm->idx; } /* Set the meter action. */ + if (!fm) { + fm = mlx5_ipool_get(priv->sh->ipool + [MLX5_IPOOL_MTR], flow->meter); + if (!fm) + return rte_flow_error_set(error, + rte_errno, + RTE_FLOW_ERROR_TYPE_ACTION, + NULL, + "meter not found " + "or invalid parameters"); + } dev_flow->dv.actions[actions_n++] = - flow->meter->mfts->meter_action; + fm->mfts->meter_action; action_flags |= MLX5_FLOW_ACTION_METER; break; case RTE_FLOW_ACTION_TYPE_SET_IPV4_DSCP: @@ -7456,7 +8246,23 @@ cnt_err: (dev, mhdr_res, dev_flow, error)) return -rte_errno; dev_flow->dv.actions[modify_action_position] = - dev_flow->dv.modify_hdr->verbs_action; + handle->dvh.modify_hdr->verbs_action; + } + if (action_flags & MLX5_FLOW_ACTION_COUNT) { + flow->counter = + flow_dv_translate_create_counter(dev, + dev_flow, count, age); + + if (!flow->counter) + return rte_flow_error_set + (error, rte_errno, + RTE_FLOW_ERROR_TYPE_ACTION, + NULL, + "cannot create counter" + " object."); + dev_flow->dv.actions[actions_n++] = + (flow_dv_counter_get_by_idx(dev, + flow->counter, NULL))->action; } break; default: @@ -7467,7 +8273,7 @@ cnt_err: modify_action_position = actions_n++; } dev_flow->dv.actions_n = actions_n; - dev_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); int item_type = items->type; @@ -7480,7 +8286,8 @@ cnt_err: break; case RTE_FLOW_ITEM_TYPE_ETH: flow_dv_translate_item_eth(match_mask, match_value, - items, tunnel); + items, tunnel, + dev_flow->dv.group); matcher.priority = MLX5_PRIORITY_MAP_L2; last_item = tunnel ? MLX5_FLOW_LAYER_INNER_L2 : MLX5_FLOW_LAYER_OUTER_L2; @@ -7488,7 +8295,8 @@ cnt_err: case RTE_FLOW_ITEM_TYPE_VLAN: flow_dv_translate_item_vlan(dev_flow, match_mask, match_value, - items, tunnel); + items, tunnel, + dev_flow->dv.group); matcher.priority = MLX5_PRIORITY_MAP_L2; last_item = tunnel ? (MLX5_FLOW_LAYER_INNER_L2 | MLX5_FLOW_LAYER_INNER_VLAN) : @@ -7499,8 +8307,8 @@ cnt_err: mlx5_flow_tunnel_ip_check(items, next_protocol, &item_flags, &tunnel); flow_dv_translate_item_ipv4(match_mask, match_value, - items, tunnel, - dev_flow->group); + items, item_flags, tunnel, + dev_flow->dv.group); matcher.priority = MLX5_PRIORITY_MAP_L3; last_item = tunnel ? MLX5_FLOW_LAYER_INNER_L3_IPV4 : MLX5_FLOW_LAYER_OUTER_L3_IPV4; @@ -7522,8 +8330,8 @@ cnt_err: mlx5_flow_tunnel_ip_check(items, next_protocol, &item_flags, &tunnel); flow_dv_translate_item_ipv6(match_mask, match_value, - items, tunnel, - dev_flow->group); + items, item_flags, tunnel, + dev_flow->dv.group); matcher.priority = MLX5_PRIORITY_MAP_L3; last_item = tunnel ? MLX5_FLOW_LAYER_INNER_L3_IPV6 : MLX5_FLOW_LAYER_OUTER_L3_IPV6; @@ -7558,6 +8366,8 @@ cnt_err: case RTE_FLOW_ITEM_TYPE_GRE: flow_dv_translate_item_gre(match_mask, match_value, items, tunnel); + matcher.priority = rss_desc->level >= 2 ? + MLX5_PRIORITY_MAP_L2 : MLX5_PRIORITY_MAP_L4; last_item = MLX5_FLOW_LAYER_GRE; break; case RTE_FLOW_ITEM_TYPE_GRE_KEY: @@ -7568,26 +8378,37 @@ cnt_err: case RTE_FLOW_ITEM_TYPE_NVGRE: flow_dv_translate_item_nvgre(match_mask, match_value, items, tunnel); + matcher.priority = rss_desc->level >= 2 ? + MLX5_PRIORITY_MAP_L2 : MLX5_PRIORITY_MAP_L4; last_item = MLX5_FLOW_LAYER_GRE; break; case RTE_FLOW_ITEM_TYPE_VXLAN: flow_dv_translate_item_vxlan(match_mask, match_value, items, tunnel); + matcher.priority = rss_desc->level >= 2 ? + MLX5_PRIORITY_MAP_L2 : MLX5_PRIORITY_MAP_L4; last_item = MLX5_FLOW_LAYER_VXLAN; break; case RTE_FLOW_ITEM_TYPE_VXLAN_GPE: - flow_dv_translate_item_vxlan(match_mask, match_value, - items, tunnel); + flow_dv_translate_item_vxlan_gpe(match_mask, + match_value, items, + tunnel); + matcher.priority = rss_desc->level >= 2 ? + MLX5_PRIORITY_MAP_L2 : MLX5_PRIORITY_MAP_L4; last_item = MLX5_FLOW_LAYER_VXLAN_GPE; break; case RTE_FLOW_ITEM_TYPE_GENEVE: flow_dv_translate_item_geneve(match_mask, match_value, items, tunnel); + matcher.priority = rss_desc->level >= 2 ? + MLX5_PRIORITY_MAP_L2 : MLX5_PRIORITY_MAP_L4; last_item = MLX5_FLOW_LAYER_GENEVE; break; case RTE_FLOW_ITEM_TYPE_MPLS: flow_dv_translate_item_mpls(match_mask, match_value, items, last_item, tunnel); + matcher.priority = rss_desc->level >= 2 ? + MLX5_PRIORITY_MAP_L2 : MLX5_PRIORITY_MAP_L4; last_item = MLX5_FLOW_LAYER_MPLS; break; case RTE_FLOW_ITEM_TYPE_MARK: @@ -7629,6 +8450,8 @@ cnt_err: case RTE_FLOW_ITEM_TYPE_GTP: flow_dv_translate_item_gtp(match_mask, match_value, items, tunnel); + matcher.priority = rss_desc->level >= 2 ? + MLX5_PRIORITY_MAP_L2 : MLX5_PRIORITY_MAP_L4; last_item = MLX5_FLOW_LAYER_GTP; break; default: @@ -7637,23 +8460,29 @@ cnt_err: item_flags |= last_item; } /* - * In case of ingress traffic when E-Switch mode is enabled, - * we have two cases where we need to set the source port manually. + * When E-Switch mode is enabled, we have two cases where we need to + * set the source port manually. * The first one, is in case of Nic steering rule, and the second is * E-Switch rule where no port_id item was found. In both cases * the source port is set according the current port in use. */ - if ((attr->ingress && !(item_flags & MLX5_FLOW_ITEM_PORT_ID)) && + if (!(item_flags & MLX5_FLOW_ITEM_PORT_ID) && (priv->representor || priv->master)) { if (flow_dv_translate_item_port_id(dev, match_mask, match_value, NULL)) return -rte_errno; } - assert(!flow_dv_check_valid_spec(matcher.mask.buf, - dev_flow->dv.value.buf)); - dev_flow->layers = item_flags; +#ifdef RTE_LIBRTE_MLX5_DEBUG + MLX5_ASSERT(!flow_dv_check_valid_spec(matcher.mask.buf, + dev_flow->dv.value.buf)); +#endif + /* + * Layers may be already initialized from prefix flow if this dev_flow + * is the suffix flow. + */ + handle->layers |= item_flags; if (action_flags & MLX5_FLOW_ACTION_RSS) - flow_dv_hashfields_set(dev_flow); + flow_dv_hashfields_set(dev_flow, rss_desc); /* Register matcher. */ matcher.crc = rte_raw_cksum((const void *)matcher.mask.buf, matcher.mask.size); @@ -7662,7 +8491,7 @@ cnt_err: /* reserved field no needs to be set to 0 here. */ tbl_key.domain = attr->transfer; tbl_key.direction = attr->egress; - tbl_key.table_id = dev_flow->group; + tbl_key.table_id = dev_flow->dv.group; if (flow_dv_matcher_register(dev, &matcher, &tbl_key, dev_flow, error)) return -rte_errno; return 0; @@ -7686,21 +8515,29 @@ static int __flow_dv_apply(struct rte_eth_dev *dev, struct rte_flow *flow, struct rte_flow_error *error) { - struct mlx5_flow_dv *dv; + struct mlx5_flow_dv_workspace *dv; + struct mlx5_flow_handle *dh; + struct mlx5_flow_handle_dv *dv_h; struct mlx5_flow *dev_flow; struct mlx5_priv *priv = dev->data->dev_private; + uint32_t handle_idx; int n; int err; + int idx; - LIST_FOREACH(dev_flow, &flow->dev_flows, next) { + for (idx = priv->flow_idx - 1; idx >= priv->flow_nested_idx; idx--) { + dev_flow = &((struct mlx5_flow *)priv->inter_flows)[idx]; dv = &dev_flow->dv; + dh = dev_flow->handle; + dv_h = &dh->dvh; n = dv->actions_n; - if (dev_flow->actions & MLX5_FLOW_ACTION_DROP) { - if (dev_flow->transfer) { + if (dh->fate_action == MLX5_FLOW_FATE_DROP) { + if (dv->transfer) { dv->actions[n++] = priv->sh->esw_drop_action; } else { - dv->hrxq = mlx5_hrxq_drop_new(dev); - if (!dv->hrxq) { + struct mlx5_hrxq *drop_hrxq; + drop_hrxq = mlx5_hrxq_drop_new(dev); + if (!drop_hrxq) { rte_flow_error_set (error, errno, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, @@ -7708,28 +8545,40 @@ __flow_dv_apply(struct rte_eth_dev *dev, struct rte_flow *flow, "cannot get drop hash queue"); goto error; } - dv->actions[n++] = dv->hrxq->action; + /* + * Drop queues will be released by the specify + * mlx5_hrxq_drop_release() function. Assign + * the special index to hrxq to mark the queue + * has been allocated. + */ + dh->rix_hrxq = UINT32_MAX; + dv->actions[n++] = drop_hrxq->action; } - } else if (dev_flow->actions & - (MLX5_FLOW_ACTION_QUEUE | MLX5_FLOW_ACTION_RSS)) { + } else if (dh->fate_action == MLX5_FLOW_FATE_QUEUE) { struct mlx5_hrxq *hrxq; - - assert(flow->rss.queue); - hrxq = mlx5_hrxq_get(dev, flow->rss.key, - MLX5_RSS_HASH_KEY_LEN, - dev_flow->hash_fields, - (*flow->rss.queue), - flow->rss.queue_num); - if (!hrxq) { - hrxq = mlx5_hrxq_new - (dev, flow->rss.key, - MLX5_RSS_HASH_KEY_LEN, - dev_flow->hash_fields, - (*flow->rss.queue), - flow->rss.queue_num, - !!(dev_flow->layers & - MLX5_FLOW_LAYER_TUNNEL)); + uint32_t hrxq_idx; + struct mlx5_flow_rss_desc *rss_desc = + &((struct mlx5_flow_rss_desc *)priv->rss_desc) + [!!priv->flow_nested_idx]; + + MLX5_ASSERT(rss_desc->queue_num); + hrxq_idx = mlx5_hrxq_get(dev, rss_desc->key, + MLX5_RSS_HASH_KEY_LEN, + dev_flow->hash_fields, + rss_desc->queue, + rss_desc->queue_num); + if (!hrxq_idx) { + hrxq_idx = mlx5_hrxq_new + (dev, rss_desc->key, + MLX5_RSS_HASH_KEY_LEN, + dev_flow->hash_fields, + rss_desc->queue, + rss_desc->queue_num, + !!(dh->layers & + MLX5_FLOW_LAYER_TUNNEL)); } + hrxq = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_HRXQ], + hrxq_idx); if (!hrxq) { rte_flow_error_set (error, rte_errno, @@ -7737,14 +8586,14 @@ __flow_dv_apply(struct rte_eth_dev *dev, struct rte_flow *flow, "cannot get hash queue"); goto error; } - dv->hrxq = hrxq; - dv->actions[n++] = dv->hrxq->action; + dh->rix_hrxq = hrxq_idx; + dv->actions[n++] = hrxq->action; } - dv->flow = - mlx5_glue->dv_create_flow(dv->matcher->matcher_object, + dh->ib_flow = + mlx5_glue->dv_create_flow(dv_h->matcher->matcher_object, (void *)&dv->value, n, dv->actions); - if (!dv->flow) { + if (!dh->ib_flow) { rte_flow_error_set(error, errno, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, @@ -7752,32 +8601,33 @@ __flow_dv_apply(struct rte_eth_dev *dev, struct rte_flow *flow, goto error; } if (priv->vmwa_context && - dev_flow->dv.vf_vlan.tag && - !dev_flow->dv.vf_vlan.created) { + dh->vf_vlan.tag && !dh->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, &dev_flow->dv.vf_vlan); + mlx5_vlan_vmwa_acquire(dev, &dh->vf_vlan); } } return 0; error: err = rte_errno; /* Save rte_errno before cleanup. */ - LIST_FOREACH(dev_flow, &flow->dev_flows, next) { - struct mlx5_flow_dv *dv = &dev_flow->dv; - if (dv->hrxq) { - if (dev_flow->actions & MLX5_FLOW_ACTION_DROP) + SILIST_FOREACH(priv->sh->ipool[MLX5_IPOOL_MLX5_FLOW], flow->dev_handles, + handle_idx, dh, next) { + /* hrxq is union, don't clear it if the flag is not set. */ + if (dh->rix_hrxq) { + if (dh->fate_action == MLX5_FLOW_FATE_DROP) { mlx5_hrxq_drop_release(dev); - else - mlx5_hrxq_release(dev, dv->hrxq); - dv->hrxq = NULL; + dh->rix_hrxq = 0; + } else if (dh->fate_action == MLX5_FLOW_FATE_QUEUE) { + mlx5_hrxq_release(dev, dh->rix_hrxq); + dh->rix_hrxq = 0; + } } - if (dev_flow->dv.vf_vlan.tag && - dev_flow->dv.vf_vlan.created) - mlx5_vlan_vmwa_release(dev, &dev_flow->dv.vf_vlan); + if (dh->vf_vlan.tag && dh->vf_vlan.created) + mlx5_vlan_vmwa_release(dev, &dh->vf_vlan); } rte_errno = err; /* Restore rte_errno. */ return -rte_errno; @@ -7788,19 +8638,19 @@ error: * * @param dev * Pointer to Ethernet device. - * @param flow - * Pointer to mlx5_flow. + * @param handle + * Pointer to mlx5_flow_handle. * * @return * 1 while a reference on it exists, 0 when freed. */ static int flow_dv_matcher_release(struct rte_eth_dev *dev, - struct mlx5_flow *flow) + struct mlx5_flow_handle *handle) { - struct mlx5_flow_dv_matcher *matcher = flow->dv.matcher; + struct mlx5_flow_dv_matcher *matcher = handle->dvh.matcher; - assert(matcher->matcher_object); + MLX5_ASSERT(matcher->matcher_object); DRV_LOG(DEBUG, "port %u matcher %p: refcnt %d--", dev->data->port_id, (void *)matcher, rte_atomic32_read(&matcher->refcnt)); @@ -7821,27 +8671,37 @@ flow_dv_matcher_release(struct rte_eth_dev *dev, /** * Release an encap/decap resource. * - * @param flow - * Pointer to mlx5_flow. + * @param dev + * Pointer to Ethernet device. + * @param handle + * Pointer to mlx5_flow_handle. * * @return * 1 while a reference on it exists, 0 when freed. */ static int -flow_dv_encap_decap_resource_release(struct mlx5_flow *flow) +flow_dv_encap_decap_resource_release(struct rte_eth_dev *dev, + struct mlx5_flow_handle *handle) { - struct mlx5_flow_dv_encap_decap_resource *cache_resource = - flow->dv.encap_decap; + struct mlx5_priv *priv = dev->data->dev_private; + uint32_t idx = handle->dvh.rix_encap_decap; + struct mlx5_flow_dv_encap_decap_resource *cache_resource; - assert(cache_resource->verbs_action); + cache_resource = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_DECAP_ENCAP], + idx); + if (!cache_resource) + return 0; + MLX5_ASSERT(cache_resource->verbs_action); DRV_LOG(DEBUG, "encap/decap resource %p: refcnt %d--", (void *)cache_resource, rte_atomic32_read(&cache_resource->refcnt)); if (rte_atomic32_dec_and_test(&cache_resource->refcnt)) { claim_zero(mlx5_glue->destroy_flow_action (cache_resource->verbs_action)); - LIST_REMOVE(cache_resource, next); - rte_free(cache_resource); + ILIST_REMOVE(priv->sh->ipool[MLX5_IPOOL_DECAP_ENCAP], + &priv->sh->encaps_decaps, idx, + cache_resource, next); + mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_DECAP_ENCAP], idx); DRV_LOG(DEBUG, "encap/decap resource %p: removed", (void *)cache_resource); return 0; @@ -7854,22 +8714,26 @@ flow_dv_encap_decap_resource_release(struct mlx5_flow *flow) * * @param dev * Pointer to Ethernet device. - * @param flow - * Pointer to mlx5_flow. + * @param handle + * Pointer to mlx5_flow_handle. * * @return * 1 while a reference on it exists, 0 when freed. */ static int flow_dv_jump_tbl_resource_release(struct rte_eth_dev *dev, - struct mlx5_flow *flow) + struct mlx5_flow_handle *handle) { - struct mlx5_flow_dv_jump_tbl_resource *cache_resource = flow->dv.jump; - struct mlx5_flow_tbl_data_entry *tbl_data = - container_of(cache_resource, - struct mlx5_flow_tbl_data_entry, jump); + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_dv_jump_tbl_resource *cache_resource; + struct mlx5_flow_tbl_data_entry *tbl_data; - assert(cache_resource->action); + tbl_data = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_JUMP], + handle->rix_jump); + if (!tbl_data) + return 0; + cache_resource = &tbl_data->jump; + MLX5_ASSERT(cache_resource->action); DRV_LOG(DEBUG, "jump table resource %p: refcnt %d--", (void *)cache_resource, rte_atomic32_read(&cache_resource->refcnt)); @@ -7888,19 +8752,19 @@ flow_dv_jump_tbl_resource_release(struct rte_eth_dev *dev, /** * Release a modify-header resource. * - * @param flow - * Pointer to mlx5_flow. + * @param handle + * Pointer to mlx5_flow_handle. * * @return * 1 while a reference on it exists, 0 when freed. */ static int -flow_dv_modify_hdr_resource_release(struct mlx5_flow *flow) +flow_dv_modify_hdr_resource_release(struct mlx5_flow_handle *handle) { struct mlx5_flow_dv_modify_hdr_resource *cache_resource = - flow->dv.modify_hdr; + handle->dvh.modify_hdr; - assert(cache_resource->verbs_action); + MLX5_ASSERT(cache_resource->verbs_action); DRV_LOG(DEBUG, "modify-header resource %p: refcnt %d--", (void *)cache_resource, rte_atomic32_read(&cache_resource->refcnt)); @@ -7919,27 +8783,37 @@ flow_dv_modify_hdr_resource_release(struct mlx5_flow *flow) /** * Release port ID action resource. * - * @param flow - * Pointer to mlx5_flow. + * @param dev + * Pointer to Ethernet device. + * @param handle + * Pointer to mlx5_flow_handle. * * @return * 1 while a reference on it exists, 0 when freed. */ static int -flow_dv_port_id_action_resource_release(struct mlx5_flow *flow) +flow_dv_port_id_action_resource_release(struct rte_eth_dev *dev, + struct mlx5_flow_handle *handle) { - struct mlx5_flow_dv_port_id_action_resource *cache_resource = - flow->dv.port_id_action; + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_dv_port_id_action_resource *cache_resource; + uint32_t idx = handle->rix_port_id_action; - assert(cache_resource->action); + cache_resource = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_PORT_ID], + idx); + if (!cache_resource) + return 0; + MLX5_ASSERT(cache_resource->action); DRV_LOG(DEBUG, "port ID action resource %p: refcnt %d--", (void *)cache_resource, rte_atomic32_read(&cache_resource->refcnt)); if (rte_atomic32_dec_and_test(&cache_resource->refcnt)) { claim_zero(mlx5_glue->destroy_flow_action (cache_resource->action)); - LIST_REMOVE(cache_resource, next); - rte_free(cache_resource); + ILIST_REMOVE(priv->sh->ipool[MLX5_IPOOL_PORT_ID], + &priv->sh->port_id_action_list, idx, + cache_resource, next); + mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_PORT_ID], idx); DRV_LOG(DEBUG, "port id action resource %p: removed", (void *)cache_resource); return 0; @@ -7950,27 +8824,37 @@ flow_dv_port_id_action_resource_release(struct mlx5_flow *flow) /** * Release push vlan action resource. * - * @param flow - * Pointer to mlx5_flow. + * @param dev + * Pointer to Ethernet device. + * @param handle + * Pointer to mlx5_flow_handle. * * @return * 1 while a reference on it exists, 0 when freed. */ static int -flow_dv_push_vlan_action_resource_release(struct mlx5_flow *flow) +flow_dv_push_vlan_action_resource_release(struct rte_eth_dev *dev, + struct mlx5_flow_handle *handle) { - struct mlx5_flow_dv_push_vlan_action_resource *cache_resource = - flow->dv.push_vlan_res; + struct mlx5_priv *priv = dev->data->dev_private; + uint32_t idx = handle->dvh.rix_push_vlan; + struct mlx5_flow_dv_push_vlan_action_resource *cache_resource; - assert(cache_resource->action); + cache_resource = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_PUSH_VLAN], + idx); + if (!cache_resource) + return 0; + MLX5_ASSERT(cache_resource->action); DRV_LOG(DEBUG, "push VLAN action resource %p: refcnt %d--", (void *)cache_resource, rte_atomic32_read(&cache_resource->refcnt)); if (rte_atomic32_dec_and_test(&cache_resource->refcnt)) { claim_zero(mlx5_glue->destroy_flow_action (cache_resource->action)); - LIST_REMOVE(cache_resource, next); - rte_free(cache_resource); + ILIST_REMOVE(priv->sh->ipool[MLX5_IPOOL_PUSH_VLAN], + &priv->sh->push_vlan_action_list, idx, + cache_resource, next); + mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_PUSH_VLAN], idx); DRV_LOG(DEBUG, "push vlan action resource %p: removed", (void *)cache_resource); return 0; @@ -7978,6 +8862,33 @@ flow_dv_push_vlan_action_resource_release(struct mlx5_flow *flow) return 1; } +/** + * Release the fate resource. + * + * @param dev + * Pointer to Ethernet device. + * @param handle + * Pointer to mlx5_flow_handle. + */ +static void +flow_dv_fate_resource_release(struct rte_eth_dev *dev, + struct mlx5_flow_handle *handle) +{ + if (!handle->rix_fate) + return; + if (handle->fate_action == MLX5_FLOW_FATE_DROP) + mlx5_hrxq_drop_release(dev); + else if (handle->fate_action == MLX5_FLOW_FATE_QUEUE) + mlx5_hrxq_release(dev, handle->rix_hrxq); + else if (handle->fate_action == MLX5_FLOW_FATE_JUMP) + flow_dv_jump_tbl_resource_release(dev, handle); + else if (handle->fate_action == MLX5_FLOW_FATE_PORT_ID) + flow_dv_port_id_action_resource_release(dev, handle); + else + DRV_LOG(DEBUG, "Incorrect fate action:%d", handle->fate_action); + handle->rix_fate = 0; +} + /** * Remove the flow from the NIC but keeps it in memory. * Lock free, (mutex should be acquired by caller). @@ -7990,27 +8901,28 @@ flow_dv_push_vlan_action_resource_release(struct mlx5_flow *flow) static void __flow_dv_remove(struct rte_eth_dev *dev, struct rte_flow *flow) { - struct mlx5_flow_dv *dv; - struct mlx5_flow *dev_flow; + struct mlx5_flow_handle *dh; + uint32_t handle_idx; + struct mlx5_priv *priv = dev->data->dev_private; if (!flow) return; - LIST_FOREACH(dev_flow, &flow->dev_flows, next) { - dv = &dev_flow->dv; - if (dv->flow) { - claim_zero(mlx5_glue->dv_destroy_flow(dv->flow)); - dv->flow = NULL; - } - if (dv->hrxq) { - if (dev_flow->actions & MLX5_FLOW_ACTION_DROP) - mlx5_hrxq_drop_release(dev); - else - mlx5_hrxq_release(dev, dv->hrxq); - dv->hrxq = NULL; + handle_idx = flow->dev_handles; + while (handle_idx) { + dh = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_MLX5_FLOW], + handle_idx); + if (!dh) + return; + if (dh->ib_flow) { + claim_zero(mlx5_glue->dv_destroy_flow(dh->ib_flow)); + dh->ib_flow = NULL; } - if (dev_flow->dv.vf_vlan.tag && - dev_flow->dv.vf_vlan.created) - mlx5_vlan_vmwa_release(dev, &dev_flow->dv.vf_vlan); + if (dh->fate_action == MLX5_FLOW_FATE_DROP || + dh->fate_action == MLX5_FLOW_FATE_QUEUE) + flow_dv_fate_resource_release(dev, dh); + if (dh->vf_vlan.tag && dh->vf_vlan.created) + mlx5_vlan_vmwa_release(dev, &dh->vf_vlan); + handle_idx = dh->next.next; } } @@ -8026,37 +8938,48 @@ __flow_dv_remove(struct rte_eth_dev *dev, struct rte_flow *flow) static void __flow_dv_destroy(struct rte_eth_dev *dev, struct rte_flow *flow) { - struct mlx5_flow *dev_flow; + struct mlx5_flow_handle *dev_handle; + struct mlx5_priv *priv = dev->data->dev_private; if (!flow) return; __flow_dv_remove(dev, flow); if (flow->counter) { flow_dv_counter_release(dev, flow->counter); - flow->counter = NULL; + flow->counter = 0; } if (flow->meter) { - mlx5_flow_meter_detach(flow->meter); - flow->meter = NULL; - } - while (!LIST_EMPTY(&flow->dev_flows)) { - dev_flow = LIST_FIRST(&flow->dev_flows); - LIST_REMOVE(dev_flow, next); - if (dev_flow->dv.matcher) - flow_dv_matcher_release(dev, dev_flow); - if (dev_flow->dv.encap_decap) - flow_dv_encap_decap_resource_release(dev_flow); - if (dev_flow->dv.modify_hdr) - flow_dv_modify_hdr_resource_release(dev_flow); - if (dev_flow->dv.jump) - flow_dv_jump_tbl_resource_release(dev, dev_flow); - if (dev_flow->dv.port_id_action) - flow_dv_port_id_action_resource_release(dev_flow); - if (dev_flow->dv.push_vlan_res) - flow_dv_push_vlan_action_resource_release(dev_flow); - if (dev_flow->dv.tag_resource) - flow_dv_tag_release(dev, dev_flow->dv.tag_resource); - rte_free(dev_flow); + struct mlx5_flow_meter *fm; + + fm = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_MTR], + flow->meter); + if (fm) + mlx5_flow_meter_detach(fm); + flow->meter = 0; + } + while (flow->dev_handles) { + uint32_t tmp_idx = flow->dev_handles; + + dev_handle = mlx5_ipool_get(priv->sh->ipool + [MLX5_IPOOL_MLX5_FLOW], tmp_idx); + if (!dev_handle) + return; + flow->dev_handles = dev_handle->next.next; + if (dev_handle->dvh.matcher) + flow_dv_matcher_release(dev, dev_handle); + if (dev_handle->dvh.rix_encap_decap) + flow_dv_encap_decap_resource_release(dev, dev_handle); + if (dev_handle->dvh.modify_hdr) + flow_dv_modify_hdr_resource_release(dev_handle); + if (dev_handle->dvh.rix_push_vlan) + flow_dv_push_vlan_action_resource_release(dev, + dev_handle); + if (dev_handle->dvh.rix_tag) + flow_dv_tag_release(dev, + dev_handle->dvh.rix_tag); + flow_dv_fate_resource_release(dev, dev_handle); + mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_MLX5_FLOW], + tmp_idx); } } @@ -8089,6 +9012,10 @@ flow_dv_query_count(struct rte_eth_dev *dev, struct rte_flow *flow, "counters are not supported"); if (flow->counter) { uint64_t pkts, bytes; + struct mlx5_flow_counter *cnt; + + cnt = flow_dv_counter_get_by_idx(dev, flow->counter, + NULL); int err = _flow_dv_query_count(dev, flow->counter, &pkts, &bytes); @@ -8098,11 +9025,11 @@ flow_dv_query_count(struct rte_eth_dev *dev, struct rte_flow *flow, NULL, "cannot read counters"); qc->hits_set = 1; qc->bytes_set = 1; - qc->hits = pkts - flow->counter->hits; - qc->bytes = bytes - flow->counter->bytes; + qc->hits = pkts - cnt->hits; + qc->bytes = bytes - cnt->bytes; if (qc->reset) { - flow->counter->hits = pkts; - flow->counter->bytes = bytes; + cnt->hits = pkts; + cnt->bytes = bytes; } return 0; } @@ -8182,8 +9109,9 @@ flow_dv_destroy_mtr_tbl(struct rte_eth_dev *dev, claim_zero(mlx5_glue->dv_destroy_flow_matcher (mtd->egress.any_matcher)); if (mtd->egress.tbl) - claim_zero(flow_dv_tbl_resource_release(dev, - mtd->egress.tbl)); + flow_dv_tbl_resource_release(dev, mtd->egress.tbl); + if (mtd->egress.sfx_tbl) + flow_dv_tbl_resource_release(dev, mtd->egress.sfx_tbl); if (mtd->ingress.color_matcher) claim_zero(mlx5_glue->dv_destroy_flow_matcher (mtd->ingress.color_matcher)); @@ -8191,8 +9119,9 @@ flow_dv_destroy_mtr_tbl(struct rte_eth_dev *dev, claim_zero(mlx5_glue->dv_destroy_flow_matcher (mtd->ingress.any_matcher)); if (mtd->ingress.tbl) - claim_zero(flow_dv_tbl_resource_release(dev, - mtd->ingress.tbl)); + flow_dv_tbl_resource_release(dev, mtd->ingress.tbl); + if (mtd->ingress.sfx_tbl) + flow_dv_tbl_resource_release(dev, mtd->ingress.sfx_tbl); if (mtd->transfer.color_matcher) claim_zero(mlx5_glue->dv_destroy_flow_matcher (mtd->transfer.color_matcher)); @@ -8200,8 +9129,9 @@ flow_dv_destroy_mtr_tbl(struct rte_eth_dev *dev, claim_zero(mlx5_glue->dv_destroy_flow_matcher (mtd->transfer.any_matcher)); if (mtd->transfer.tbl) - claim_zero(flow_dv_tbl_resource_release(dev, - mtd->transfer.tbl)); + flow_dv_tbl_resource_release(dev, mtd->transfer.tbl); + if (mtd->transfer.sfx_tbl) + flow_dv_tbl_resource_release(dev, mtd->transfer.sfx_tbl); if (mtd->drop_actn) claim_zero(mlx5_glue->destroy_flow_action(mtd->drop_actn)); rte_free(mtd); @@ -8249,31 +9179,16 @@ flow_dv_prepare_mtr_tables(struct rte_eth_dev *dev, .match_mask = (void *)&mask, }; void *actions[METER_ACTIONS]; - struct mlx5_flow_tbl_resource **sfx_tbl; struct mlx5_meter_domain_info *dtb; struct rte_flow_error error; int i = 0; - if (transfer) { - sfx_tbl = &sh->fdb_mtr_sfx_tbl; + if (transfer) dtb = &mtb->transfer; - } else if (egress) { - sfx_tbl = &sh->tx_mtr_sfx_tbl; + else if (egress) dtb = &mtb->egress; - } else { - sfx_tbl = &sh->rx_mtr_sfx_tbl; + else dtb = &mtb->ingress; - } - /* If the suffix table in missing, create it. */ - if (!(*sfx_tbl)) { - *sfx_tbl = flow_dv_tbl_resource_get(dev, - MLX5_FLOW_TABLE_LEVEL_SUFFIX, - egress, transfer, &error); - if (!(*sfx_tbl)) { - DRV_LOG(ERR, "Failed to create meter suffix table."); - return -1; - } - } /* Create the meter table with METER level. */ dtb->tbl = flow_dv_tbl_resource_get(dev, MLX5_FLOW_TABLE_LEVEL_METER, egress, transfer, &error); @@ -8281,6 +9196,14 @@ flow_dv_prepare_mtr_tables(struct rte_eth_dev *dev, DRV_LOG(ERR, "Failed to create meter policer table."); return -1; } + /* Create the meter suffix table with SUFFIX level. */ + dtb->sfx_tbl = flow_dv_tbl_resource_get(dev, + MLX5_FLOW_TABLE_LEVEL_SUFFIX, + egress, transfer, &error); + if (!dtb->sfx_tbl) { + DRV_LOG(ERR, "Failed to create meter suffix table."); + return -1; + } /* Create matchers, Any and Color. */ dv_attr.priority = 3; dv_attr.match_criteria_enable = 0; @@ -8296,7 +9219,7 @@ flow_dv_prepare_mtr_tables(struct rte_eth_dev *dev, dv_attr.match_criteria_enable = 1 << MLX5_MATCH_CRITERIA_ENABLE_MISC2_BIT; flow_dv_match_meta_reg(mask.buf, value.buf, color_reg_c_idx, - rte_col_2_mlx5_col(RTE_COLORS), UINT32_MAX); + rte_col_2_mlx5_col(RTE_COLORS), UINT8_MAX); dtb->color_matcher = mlx5_glue->dv_create_flow_matcher(sh->ctx, &dv_attr, dtb->tbl->obj); @@ -8352,9 +9275,12 @@ flow_dv_create_mtr_tbl(struct rte_eth_dev *dev, } /* Create meter count actions */ for (i = 0; i <= RTE_MTR_DROPPED; i++) { + struct mlx5_flow_counter *cnt; if (!fm->policer_stats.cnt[i]) continue; - mtb->count_actns[i] = fm->policer_stats.cnt[i]->action; + cnt = flow_dv_counter_get_by_idx(dev, + fm->policer_stats.cnt[i], NULL); + mtb->count_actns[i] = cnt->action; } /* Create drop action. */ mtb->drop_actn = mlx5_glue->dr_create_flow_action_drop(); @@ -8451,8 +9377,6 @@ flow_dv_destroy_policer_rules(struct rte_eth_dev *dev __rte_unused, * Pointer to flow meter structure. * @param[in] mtb * Pointer to DV meter table set. - * @param[in] sfx_tb - * Pointer to suffix table. * @param[in] mtr_reg_c * Color match REG_C. * @@ -8462,7 +9386,6 @@ flow_dv_destroy_policer_rules(struct rte_eth_dev *dev __rte_unused, static int flow_dv_create_policer_forward_rule(struct mlx5_flow_meter *fm, struct mlx5_meter_domain_info *dtb, - struct mlx5_flow_tbl_resource *sfx_tb, uint8_t mtr_reg_c) { struct mlx5_flow_dv_match_params matcher = { @@ -8476,12 +9399,10 @@ flow_dv_create_policer_forward_rule(struct mlx5_flow_meter *fm, int i; /* Create jump action. */ - if (!sfx_tb) - return -1; if (!dtb->jump_actn) dtb->jump_actn = mlx5_glue->dr_create_flow_action_dest_flow_tbl - (sfx_tb->obj); + (dtb->sfx_tbl->obj); if (!dtb->jump_actn) { DRV_LOG(ERR, "Failed to create policer jump action."); goto error; @@ -8490,10 +9411,10 @@ flow_dv_create_policer_forward_rule(struct mlx5_flow_meter *fm, int j = 0; flow_dv_match_meta_reg(matcher.buf, value.buf, mtr_reg_c, - rte_col_2_mlx5_col(i), UINT32_MAX); + rte_col_2_mlx5_col(i), UINT8_MAX); if (mtb->count_actns[i]) actions[j++] = mtb->count_actns[i]; - if (fm->params.action[i] == MTR_POLICER_ACTION_DROP) + if (fm->action[i] == MTR_POLICER_ACTION_DROP) actions[j++] = mtb->drop_actn; else actions[j++] = dtb->jump_actn; @@ -8536,7 +9457,6 @@ flow_dv_create_policer_rules(struct rte_eth_dev *dev, if (attr->egress) { ret = flow_dv_create_policer_forward_rule(fm, &mtb->egress, - priv->sh->tx_mtr_sfx_tbl, priv->mtr_color_reg); if (ret) { DRV_LOG(ERR, "Failed to create egress policer."); @@ -8545,7 +9465,6 @@ flow_dv_create_policer_rules(struct rte_eth_dev *dev, } if (attr->ingress) { ret = flow_dv_create_policer_forward_rule(fm, &mtb->ingress, - priv->sh->rx_mtr_sfx_tbl, priv->mtr_color_reg); if (ret) { DRV_LOG(ERR, "Failed to create ingress policer."); @@ -8554,7 +9473,6 @@ flow_dv_create_policer_rules(struct rte_eth_dev *dev, } if (attr->transfer) { ret = flow_dv_create_policer_forward_rule(fm, &mtb->transfer, - priv->sh->fdb_mtr_sfx_tbl, priv->mtr_color_reg); if (ret) { DRV_LOG(ERR, "Failed to create transfer policer."); @@ -8573,7 +9491,7 @@ error: * @param[in] dev * Pointer to the Ethernet device structure. * @param[in] cnt - * Pointer to the flow counter. + * Index to the flow counter. * @param[in] clear * Set to clear the counter statistics. * @param[out] pkts @@ -8585,19 +9503,21 @@ error: * 0 on success, otherwise return -1. */ static int -flow_dv_counter_query(struct rte_eth_dev *dev, - struct mlx5_flow_counter *cnt, bool clear, +flow_dv_counter_query(struct rte_eth_dev *dev, uint32_t counter, bool clear, uint64_t *pkts, uint64_t *bytes) { struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_counter *cnt; uint64_t inn_pkts, inn_bytes; int ret; if (!priv->config.devx) return -1; - ret = _flow_dv_query_count(dev, cnt, &inn_pkts, &inn_bytes); + + ret = _flow_dv_query_count(dev, counter, &inn_pkts, &inn_bytes); if (ret) return -1; + cnt = flow_dv_counter_get_by_idx(dev, counter, NULL); *pkts = inn_pkts - cnt->hits; *bytes = inn_bytes - cnt->bytes; if (clear) { @@ -8607,6 +9527,60 @@ flow_dv_counter_query(struct rte_eth_dev *dev, return 0; } +/** + * Get aged-out flows. + * + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] context + * The address of an array of pointers to the aged-out flows contexts. + * @param[in] nb_contexts + * The length of context array pointers. + * @param[out] error + * Perform verbose error reporting if not NULL. Initialized in case of + * error only. + * + * @return + * how many contexts get in success, otherwise negative errno value. + * if nb_contexts is 0, return the amount of all aged contexts. + * if nb_contexts is not 0 , return the amount of aged flows reported + * in the context array. + * @note: only stub for now + */ +static int +flow_get_aged_flows(struct rte_eth_dev *dev, + void **context, + uint32_t nb_contexts, + struct rte_flow_error *error) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_age_info *age_info; + struct mlx5_age_param *age_param; + struct mlx5_flow_counter *counter; + int nb_flows = 0; + + if (nb_contexts && !context) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, + NULL, + "Should assign at least one flow or" + " context to get if nb_contexts != 0"); + age_info = GET_PORT_AGE_INFO(priv); + rte_spinlock_lock(&age_info->aged_sl); + TAILQ_FOREACH(counter, &age_info->aged_counters, next) { + nb_flows++; + if (nb_contexts) { + age_param = MLX5_CNT_TO_AGE(counter); + context[nb_flows - 1] = age_param->context; + if (!(--nb_contexts)) + break; + } + } + rte_spinlock_unlock(&age_info->aged_sl); + MLX5_AGE_SET(age_info, MLX5_AGE_TRIGGER); + return nb_flows; +} + /* * Mutex-protected thunk to lock-free __flow_dv_translate(). */ @@ -8667,13 +9641,13 @@ flow_dv_destroy(struct rte_eth_dev *dev, struct rte_flow *flow) /* * Mutex-protected thunk to lock-free flow_dv_counter_alloc(). */ -static struct mlx5_flow_counter * +static uint32_t flow_dv_counter_allocate(struct rte_eth_dev *dev) { - struct mlx5_flow_counter *cnt; + uint32_t cnt; flow_dv_shared_lock(dev); - cnt = flow_dv_counter_alloc(dev, 0, 0, 1); + cnt = flow_dv_counter_alloc(dev, 0, 0, 1, 0); flow_dv_shared_unlock(dev); return cnt; } @@ -8682,7 +9656,7 @@ flow_dv_counter_allocate(struct rte_eth_dev *dev) * Mutex-protected thunk to lock-free flow_dv_counter_release(). */ static void -flow_dv_counter_free(struct rte_eth_dev *dev, struct mlx5_flow_counter *cnt) +flow_dv_counter_free(struct rte_eth_dev *dev, uint32_t cnt) { flow_dv_shared_lock(dev); flow_dv_counter_release(dev, cnt); @@ -8704,6 +9678,7 @@ const struct mlx5_flow_driver_ops mlx5_flow_dv_drv_ops = { .counter_alloc = flow_dv_counter_allocate, .counter_free = flow_dv_counter_free, .counter_query = flow_dv_counter_query, + .get_aged_flows = flow_get_aged_flows, }; #endif /* HAVE_IBV_FLOW_DV_SUPPORT */