From: Wei Zhao Date: Wed, 10 Jan 2018 02:10:26 +0000 (+0800) Subject: net/i40e: move RSS to flow API X-Git-Tag: spdx-start~327 X-Git-Url: http://git.droids-corp.org/?a=commitdiff_plain;h=ecad87d22383e4cfe4b6838342c2de6e18b05c28;p=dpdk.git net/i40e: move RSS to flow API Rte_flow was defined to include RSS, this patch moves i40e existing RSS to rte_flow. The old RSS configuration is kept as it was, and can be deprecated in the future. Signed-off-by: Wei Zhao Acked-by: Qi Zhang --- diff --git a/doc/guides/rel_notes/release_18_02.rst b/doc/guides/rel_notes/release_18_02.rst index 6c755bf777..24a5080cf9 100644 --- a/doc/guides/rel_notes/release_18_02.rst +++ b/doc/guides/rel_notes/release_18_02.rst @@ -61,6 +61,13 @@ New Features 1, 2, 4, 8 or 16. If no such parameter is configured, the number of queues per VF is 4 by default. +* **Added the i40e ethernet driver to support RSS with flow API.** + + Rte_flow actually defined to include RSS, but till now, RSS is out of + rte_flow. This patch is to support i40e NIC with existing RSS + configuration using rte_flow API.It also enable queue region configuration + using flow API for i40e. + * **Added NVGRE and UDP tunnels support in Solarflare network PMD.** Added support for NVGRE, VXLAN and GENEVE tunnels. diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c index 83bcec6a9b..f8abd5bf75 100644 --- a/drivers/net/i40e/i40e_ethdev.c +++ b/drivers/net/i40e/i40e_ethdev.c @@ -1313,6 +1313,10 @@ eth_i40e_dev_init(struct rte_eth_dev *dev) /* initialize queue region configuration */ i40e_init_queue_region_conf(dev); + /* initialize rss configuration from rte_flow */ + memset(&pf->rss_info, 0, + sizeof(struct i40e_rte_flow_rss_conf)); + return 0; err_init_fdir_filter_list: @@ -11123,12 +11127,23 @@ i40e_tunnel_filter_restore(struct i40e_pf *pf) } } +/* Restore rss filter */ +static inline void +i40e_rss_filter_restore(struct i40e_pf *pf) +{ + struct i40e_rte_flow_rss_conf *conf = + &pf->rss_info; + if (conf->num) + i40e_config_rss_filter(pf, conf, TRUE); +} + static void i40e_filter_restore(struct i40e_pf *pf) { i40e_ethertype_filter_restore(pf); i40e_tunnel_filter_restore(pf); i40e_fdir_filter_restore(pf); + i40e_rss_filter_restore(pf); } static bool @@ -11583,6 +11598,82 @@ i40e_cloud_filter_qinq_create(struct i40e_pf *pf) return ret; } +int +i40e_config_rss_filter(struct i40e_pf *pf, + struct i40e_rte_flow_rss_conf *conf, bool add) +{ + struct i40e_hw *hw = I40E_PF_TO_HW(pf); + uint32_t i, lut = 0; + uint16_t j, num; + struct rte_eth_rss_conf rss_conf = conf->rss_conf; + struct i40e_rte_flow_rss_conf *rss_info = &pf->rss_info; + + if (!add) { + if (memcmp(conf, rss_info, + sizeof(struct i40e_rte_flow_rss_conf)) == 0) { + i40e_pf_disable_rss(pf); + memset(rss_info, 0, + sizeof(struct i40e_rte_flow_rss_conf)); + return 0; + } + return -EINVAL; + } + + if (rss_info->num) + return -EINVAL; + + /* If both VMDQ and RSS enabled, not all of PF queues are configured. + * It's necessary to calculate the actual PF queues that are configured. + */ + if (pf->dev_data->dev_conf.rxmode.mq_mode & ETH_MQ_RX_VMDQ_FLAG) + num = i40e_pf_calc_configured_queues_num(pf); + else + num = pf->dev_data->nb_rx_queues; + + num = RTE_MIN(num, conf->num); + PMD_DRV_LOG(INFO, "Max of contiguous %u PF queues are configured", + num); + + if (num == 0) { + PMD_DRV_LOG(ERR, "No PF queues are configured to enable RSS"); + return -ENOTSUP; + } + + /* Fill in redirection table */ + for (i = 0, j = 0; i < hw->func_caps.rss_table_size; i++, j++) { + if (j == num) + j = 0; + lut = (lut << 8) | (conf->queue[j] & ((0x1 << + hw->func_caps.rss_table_entry_width) - 1)); + if ((i & 3) == 3) + I40E_WRITE_REG(hw, I40E_PFQF_HLUT(i >> 2), lut); + } + + if ((rss_conf.rss_hf & pf->adapter->flow_types_mask) == 0) { + i40e_pf_disable_rss(pf); + return 0; + } + if (rss_conf.rss_key == NULL || rss_conf.rss_key_len < + (I40E_PFQF_HKEY_MAX_INDEX + 1) * sizeof(uint32_t)) { + /* Random default keys */ + static uint32_t rss_key_default[] = {0x6b793944, + 0x23504cb5, 0x5bea75b6, 0x309f4f12, 0x3dc0a2b8, + 0x024ddcdf, 0x339b8ca0, 0x4c4af64a, 0x34fac605, + 0x55d85839, 0x3a58997d, 0x2ec938e1, 0x66031581}; + + rss_conf.rss_key = (uint8_t *)rss_key_default; + rss_conf.rss_key_len = (I40E_PFQF_HKEY_MAX_INDEX + 1) * + sizeof(uint32_t); + } + + return i40e_hw_rss_hash_set(pf, &rss_conf); + + rte_memcpy(rss_info, + conf, sizeof(struct i40e_rte_flow_rss_conf)); + + return 0; +} + RTE_INIT(i40e_init_log); static void i40e_init_log(void) diff --git a/drivers/net/i40e/i40e_ethdev.h b/drivers/net/i40e/i40e_ethdev.h index 84a5133a4b..69ea6c189a 100644 --- a/drivers/net/i40e/i40e_ethdev.h +++ b/drivers/net/i40e/i40e_ethdev.h @@ -867,6 +867,13 @@ struct i40e_customized_pctype { bool valid; /* Check if it's valid */ }; +struct i40e_rte_flow_rss_conf { + struct rte_eth_rss_conf rss_conf; /**< RSS parameters. */ + uint16_t queue_region_conf; /**< Queue region config flag */ + uint16_t num; /**< Number of entries in queue[]. */ + uint16_t queue[I40E_MAX_Q_PER_TC]; /**< Queues indices to use. */ +}; + /* * Structure to store private data specific for PF instance. */ @@ -921,6 +928,7 @@ struct i40e_pf { struct i40e_fdir_info fdir; /* flow director info */ struct i40e_ethertype_rule ethertype; /* Ethertype filter rule */ struct i40e_tunnel_rule tunnel; /* Tunnel filter rule */ + struct i40e_rte_flow_rss_conf rss_info; /* rss info */ struct i40e_queue_regions queue_region; /* queue region info */ struct i40e_fc_conf fc_conf; /* Flow control conf */ struct i40e_mirror_rule_list mirror_list; @@ -1047,6 +1055,7 @@ union i40e_filter_t { struct i40e_fdir_filter_conf fdir_filter; struct rte_eth_tunnel_filter_conf tunnel_filter; struct i40e_tunnel_filter_conf consistent_tunnel_filter; + struct i40e_rte_flow_rss_conf rss_conf; }; typedef int (*parse_filter_t)(struct rte_eth_dev *dev, @@ -1177,6 +1186,8 @@ void i40e_init_queue_region_conf(struct rte_eth_dev *dev); void i40e_flex_payload_reg_set_default(struct i40e_hw *hw); int i40e_set_rss_key(struct i40e_vsi *vsi, uint8_t *key, uint8_t key_len); int i40e_set_rss_lut(struct i40e_vsi *vsi, uint8_t *lut, uint16_t lut_size); +int i40e_config_rss_filter(struct i40e_pf *pf, + struct i40e_rte_flow_rss_conf *conf, bool add); #define I40E_DEV_TO_PCI(eth_dev) \ RTE_DEV_TO_PCI((eth_dev)->device) diff --git a/drivers/net/i40e/i40e_flow.c b/drivers/net/i40e/i40e_flow.c index 6feb7aa4df..cd9a9b64b3 100644 --- a/drivers/net/i40e/i40e_flow.c +++ b/drivers/net/i40e/i40e_flow.c @@ -110,6 +110,8 @@ static int i40e_flow_flush_fdir_filter(struct i40e_pf *pf); static int i40e_flow_flush_ethertype_filter(struct i40e_pf *pf); static int i40e_flow_flush_tunnel_filter(struct i40e_pf *pf); static int +i40e_flow_flush_rss_filter(struct rte_eth_dev *dev); +static int i40e_flow_parse_qinq_filter(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, const struct rte_flow_item pattern[], @@ -4108,6 +4110,316 @@ i40e_flow_parse_qinq_filter(struct rte_eth_dev *dev, return ret; } +/** + * This function is used to do configuration i40e existing RSS with rte_flow. + * It also enable queue region configuration using flow API for i40e. + * pattern can be used indicate what parameters will be include in flow, + * like user_priority or flowtype for queue region or HASH function for RSS. + * Action is used to transmit parameter like queue index and HASH + * function for RSS, or flowtype for queue region configuration. + * For example: + * pattern: + * Case 1: only ETH, indicate flowtype for queue region will be parsed. + * Case 2: only VLAN, indicate user_priority for queue region will be parsed. + * Case 3: none, indicate RSS related will be parsed in action. + * Any pattern other the ETH or VLAN will be treated as invalid except END. + * So, pattern choice is depened on the purpose of configuration of + * that flow. + * action: + * action RSS will be uaed to transmit valid parameter with + * struct rte_flow_action_rss for all the 3 case. + */ +static int +i40e_flow_parse_rss_pattern(__rte_unused struct rte_eth_dev *dev, + const struct rte_flow_item *pattern, + struct rte_flow_error *error, + uint8_t *action_flag, + struct i40e_queue_regions *info) +{ + const struct rte_flow_item_vlan *vlan_spec, *vlan_mask; + const struct rte_flow_item *item = pattern; + enum rte_flow_item_type item_type; + + if (item->type == RTE_FLOW_ITEM_TYPE_END) + return 0; + + for (; item->type != RTE_FLOW_ITEM_TYPE_END; item++) { + if (item->last) { + rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ITEM, + item, + "Not support range"); + return -rte_errno; + } + item_type = item->type; + switch (item_type) { + case RTE_FLOW_ITEM_TYPE_ETH: + *action_flag = 1; + break; + case RTE_FLOW_ITEM_TYPE_VLAN: + vlan_spec = + (const struct rte_flow_item_vlan *)item->spec; + vlan_mask = + (const struct rte_flow_item_vlan *)item->mask; + if (vlan_spec && vlan_mask) { + if (vlan_mask->tci == + rte_cpu_to_be_16(I40E_TCI_MASK)) { + info->region[0].user_priority[0] = + (vlan_spec->tci >> 13) & 0x7; + info->region[0].user_priority_num = 1; + info->queue_region_number = 1; + *action_flag = 0; + } + } + break; + default: + rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ITEM, + item, + "Not support range"); + return -rte_errno; + } + } + + return 0; +} + +static int +i40e_flow_parse_rss_action(struct rte_eth_dev *dev, + const struct rte_flow_action *actions, + struct rte_flow_error *error, + uint8_t *action_flag, + struct i40e_queue_regions *conf_info, + union i40e_filter_t *filter) +{ + const struct rte_flow_action *act; + const struct rte_flow_action_rss *rss; + struct i40e_pf *pf = I40E_DEV_PRIVATE_TO_PF(dev->data->dev_private); + struct i40e_queue_regions *info = &pf->queue_region; + struct i40e_rte_flow_rss_conf *rss_config = + &filter->rss_conf; + struct i40e_rte_flow_rss_conf *rss_info = &pf->rss_info; + uint16_t i, j, n, tmp; + uint32_t index = 0; + + NEXT_ITEM_OF_ACTION(act, actions, index); + rss = (const struct rte_flow_action_rss *)act->conf; + + /** + * rss only supports forwarding, + * check if the first not void action is RSS. + */ + if (act->type != RTE_FLOW_ACTION_TYPE_RSS) { + memset(rss_config, 0, sizeof(struct i40e_rte_flow_rss_conf)); + rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, + act, "Not supported action."); + return -rte_errno; + } + + if (action_flag) { + for (n = 0; n < 64; n++) { + if (rss->rss_conf->rss_hf & (1 << n)) { + conf_info->region[0].hw_flowtype[0] = n; + conf_info->region[0].flowtype_num = 1; + conf_info->queue_region_number = 1; + break; + } + } + } + + for (n = 0; n < conf_info->queue_region_number; n++) { + if (conf_info->region[n].user_priority_num || + conf_info->region[n].flowtype_num) { + if (!((rte_is_power_of_2(rss->num)) && + rss->num <= 64)) { + PMD_DRV_LOG(ERR, "The region sizes should be any of the following values: 1, 2, 4, 8, 16, 32, 64 as long as the " + "total number of queues do not exceed the VSI allocation"); + return -rte_errno; + } + + if (conf_info->region[n].user_priority[n] >= + I40E_MAX_USER_PRIORITY) { + PMD_DRV_LOG(ERR, "the user priority max index is 7"); + return -rte_errno; + } + + if (conf_info->region[n].hw_flowtype[n] >= + I40E_FILTER_PCTYPE_MAX) { + PMD_DRV_LOG(ERR, "the hw_flowtype or PCTYPE max index is 63"); + return -rte_errno; + } + + if (rss_info->num < rss->num || + rss_info->queue[0] < rss->queue[0] || + (rss->queue[0] + rss->num > + rss_info->num + rss_info->queue[0])) { + rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, + act, + "no valid queues"); + return -rte_errno; + } + + for (i = 0; i < info->queue_region_number; i++) { + if (info->region[i].queue_num == rss->num && + info->region[i].queue_start_index == + rss->queue[0]) + break; + } + + if (i == info->queue_region_number) { + if (i > I40E_REGION_MAX_INDEX) { + PMD_DRV_LOG(ERR, "the queue region max index is 7"); + return -rte_errno; + } + + info->region[i].queue_num = + rss->num; + info->region[i].queue_start_index = + rss->queue[0]; + info->region[i].region_id = + info->queue_region_number; + + j = info->region[i].user_priority_num; + tmp = conf_info->region[n].user_priority[0]; + if (conf_info->region[n].user_priority_num) { + info->region[i].user_priority[j] = tmp; + info->region[i].user_priority_num++; + } + + j = info->region[i].flowtype_num; + tmp = conf_info->region[n].hw_flowtype[0]; + if (conf_info->region[n].flowtype_num) { + info->region[i].hw_flowtype[j] = tmp; + info->region[i].flowtype_num++; + } + info->queue_region_number++; + } else { + j = info->region[i].user_priority_num; + tmp = conf_info->region[n].user_priority[0]; + if (conf_info->region[n].user_priority_num) { + info->region[i].user_priority[j] = tmp; + info->region[i].user_priority_num++; + } + + j = info->region[i].flowtype_num; + tmp = conf_info->region[n].hw_flowtype[0]; + if (conf_info->region[n].flowtype_num) { + info->region[i].hw_flowtype[j] = tmp; + info->region[i].flowtype_num++; + } + } + } + + rss_config->queue_region_conf = TRUE; + return 0; + } + + if (!rss || !rss->num) { + rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, + act, + "no valid queues"); + return -rte_errno; + } + + for (n = 0; n < rss->num; n++) { + if (rss->queue[n] >= dev->data->nb_rx_queues) { + rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, + act, + "queue id > max number of queues"); + return -rte_errno; + } + } + if (rss->rss_conf) + rss_config->rss_conf = *rss->rss_conf; + else + rss_config->rss_conf.rss_hf = + pf->adapter->flow_types_mask; + + for (n = 0; n < rss->num; ++n) + rss_config->queue[n] = rss->queue[n]; + rss_config->num = rss->num; + index++; + + /* check if the next not void action is END */ + NEXT_ITEM_OF_ACTION(act, actions, index); + if (act->type != RTE_FLOW_ACTION_TYPE_END) { + memset(rss_config, 0, sizeof(struct i40e_rte_flow_rss_conf)); + rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, + act, "Not supported action."); + return -rte_errno; + } + rss_config->queue_region_conf = FALSE; + + return 0; +} + +static int +i40e_parse_rss_filter(struct rte_eth_dev *dev, + const struct rte_flow_attr *attr, + const struct rte_flow_item pattern[], + const struct rte_flow_action actions[], + union i40e_filter_t *filter, + struct rte_flow_error *error) +{ + int ret; + struct i40e_queue_regions info; + uint8_t action_flag = 0; + + memset(&info, 0, sizeof(struct i40e_queue_regions)); + + ret = i40e_flow_parse_rss_pattern(dev, pattern, + error, &action_flag, &info); + if (ret) + return ret; + + ret = i40e_flow_parse_rss_action(dev, actions, error, + &action_flag, &info, filter); + if (ret) + return ret; + + ret = i40e_flow_parse_attr(attr, error); + if (ret) + return ret; + + cons_filter_type = RTE_ETH_FILTER_HASH; + + return 0; +} + +static int +i40e_config_rss_filter_set(struct rte_eth_dev *dev, + struct i40e_rte_flow_rss_conf *conf) +{ + struct i40e_pf *pf = I40E_DEV_PRIVATE_TO_PF(dev->data->dev_private); + struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private); + + if (conf->queue_region_conf) { + i40e_flush_queue_region_all_conf(dev, hw, pf, 1); + conf->queue_region_conf = 0; + } else { + i40e_config_rss_filter(pf, conf, 1); + } + return 0; +} + +static int +i40e_config_rss_filter_del(struct rte_eth_dev *dev, + struct i40e_rte_flow_rss_conf *conf) +{ + struct i40e_pf *pf = I40E_DEV_PRIVATE_TO_PF(dev->data->dev_private); + struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private); + + i40e_flush_queue_region_all_conf(dev, hw, pf, 0); + + i40e_config_rss_filter(pf, conf, 0); + return 0; +} + static int i40e_flow_validate(struct rte_eth_dev *dev, const struct rte_flow_attr *attr, @@ -4144,6 +4456,17 @@ i40e_flow_validate(struct rte_eth_dev *dev, memset(&cons_filter, 0, sizeof(cons_filter)); + /* Get the non-void item of action */ + while ((actions + i)->type == RTE_FLOW_ACTION_TYPE_VOID) + i++; + + if ((actions + i)->type == RTE_FLOW_ACTION_TYPE_RSS) { + ret = i40e_parse_rss_filter(dev, attr, pattern, + actions, &cons_filter, error); + return ret; + } + + i = 0; /* Get the non-void item number of pattern */ while ((pattern + i)->type != RTE_FLOW_ITEM_TYPE_END) { if ((pattern + i)->type != RTE_FLOW_ITEM_TYPE_VOID) @@ -4231,6 +4554,11 @@ i40e_flow_create(struct rte_eth_dev *dev, flow->rule = TAILQ_LAST(&pf->tunnel.tunnel_list, i40e_tunnel_filter_list); break; + case RTE_ETH_FILTER_HASH: + ret = i40e_config_rss_filter_set(dev, + &cons_filter.rss_conf); + flow->rule = &pf->rss_info; + break; default: goto free_flow; } @@ -4269,6 +4597,10 @@ i40e_flow_destroy(struct rte_eth_dev *dev, ret = i40e_flow_add_del_fdir_filter(dev, &((struct i40e_fdir_filter *)flow->rule)->fdir, 0); break; + case RTE_ETH_FILTER_HASH: + ret = i40e_config_rss_filter_del(dev, + (struct i40e_rte_flow_rss_conf *)flow->rule); + break; default: PMD_DRV_LOG(WARNING, "Filter type (%d) not supported", filter_type); @@ -4411,6 +4743,14 @@ i40e_flow_flush(struct rte_eth_dev *dev, struct rte_flow_error *error) return -rte_errno; } + ret = i40e_flow_flush_rss_filter(dev); + if (ret) { + rte_flow_error_set(error, -ret, + RTE_FLOW_ERROR_TYPE_HANDLE, NULL, + "Failed to flush rss flows."); + return -rte_errno; + } + return ret; } @@ -4506,3 +4846,19 @@ i40e_flow_flush_tunnel_filter(struct i40e_pf *pf) return ret; } + +/* remove the rss filter */ +static int +i40e_flow_flush_rss_filter(struct rte_eth_dev *dev) +{ + struct i40e_pf *pf = I40E_DEV_PRIVATE_TO_PF(dev->data->dev_private); + struct i40e_rte_flow_rss_conf *rss_info = &pf->rss_info; + struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private); + int32_t ret = -EINVAL; + + ret = i40e_flush_queue_region_all_conf(dev, hw, pf, 0); + + if (rss_info->num) + ret = i40e_config_rss_filter(pf, rss_info, FALSE); + return ret; +}