diff options
Diffstat (limited to 'drivers/net/ethernet/mellanox/mlx5/core/en')
21 files changed, 1220 insertions, 250 deletions
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/en/Makefile deleted file mode 100644 index c78512eed8d7..000000000000 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/Makefile +++ /dev/null @@ -1,2 +0,0 @@ -# SPDX-License-Identifier: GPL-2.0-only -subdir-ccflags-y += -I$(src)/.. diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h b/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h index be5961ff24cc..68d593074f6c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h @@ -10,11 +10,14 @@ enum { }; struct mlx5e_tc_table { + /* protects flow table */ + struct mutex t_lock; struct mlx5_flow_table *t; struct rhashtable ht; - DECLARE_HASHTABLE(mod_hdr_tbl, 8); + struct mod_hdr_tbl mod_hdr; + struct mutex hairpin_tbl_lock; /* protects hairpin_tbl */ DECLARE_HASHTABLE(hairpin_tbl, 8); struct notifier_block netdevice_nb; @@ -92,9 +95,15 @@ struct mlx5e_tirc_config { enum mlx5e_tunnel_types { MLX5E_TT_IPV4_GRE, MLX5E_TT_IPV6_GRE, + MLX5E_TT_IPV4_IPIP, + MLX5E_TT_IPV6_IPIP, + MLX5E_TT_IPV4_IPV6, + MLX5E_TT_IPV6_IPV6, MLX5E_NUM_TUNNEL_TT, }; +bool mlx5e_tunnel_inner_ft_supported(struct mlx5_core_dev *mdev); + /* L3/L4 traffic type classifier */ struct mlx5e_ttc_table { struct mlx5e_flow_table ft; @@ -132,12 +141,17 @@ struct mlx5e_ethtool_steering { void mlx5e_ethtool_init_steering(struct mlx5e_priv *priv); void mlx5e_ethtool_cleanup_steering(struct mlx5e_priv *priv); -int mlx5e_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd); -int mlx5e_get_rxnfc(struct net_device *dev, - struct ethtool_rxnfc *info, u32 *rule_locs); +int mlx5e_ethtool_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd); +int mlx5e_ethtool_get_rxnfc(struct net_device *dev, + struct ethtool_rxnfc *info, u32 *rule_locs); #else static inline void mlx5e_ethtool_init_steering(struct mlx5e_priv *priv) { } static inline void mlx5e_ethtool_cleanup_steering(struct mlx5e_priv *priv) { } +static inline int mlx5e_ethtool_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd) +{ return -EOPNOTSUPP; } +static inline int mlx5e_ethtool_get_rxnfc(struct net_device *dev, + struct ethtool_rxnfc *info, u32 *rule_locs) +{ return -EOPNOTSUPP; } #endif /* CONFIG_MLX5_EN_RXNFC */ #ifdef CONFIG_MLX5_EN_ARFS @@ -224,5 +238,8 @@ void mlx5e_disable_cvlan_filter(struct mlx5e_priv *priv); int mlx5e_create_flow_steering(struct mlx5e_priv *priv); void mlx5e_destroy_flow_steering(struct mlx5e_priv *priv); +bool mlx5e_tunnel_proto_supported(struct mlx5_core_dev *mdev, u8 proto_type); +bool mlx5e_any_tunnel_proto_supported(struct mlx5_core_dev *mdev); + #endif /* __MLX5E_FLOW_STEER_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/health.c b/drivers/net/ethernet/mellanox/mlx5/core/en/health.c new file mode 100644 index 000000000000..1d6b58860da6 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/health.c @@ -0,0 +1,205 @@ +// SPDX-License-Identifier: GPL-2.0 +// Copyright (c) 2019 Mellanox Technologies. + +#include "health.h" +#include "lib/eq.h" + +int mlx5e_reporter_named_obj_nest_start(struct devlink_fmsg *fmsg, char *name) +{ + int err; + + err = devlink_fmsg_pair_nest_start(fmsg, name); + if (err) + return err; + + err = devlink_fmsg_obj_nest_start(fmsg); + if (err) + return err; + + return 0; +} + +int mlx5e_reporter_named_obj_nest_end(struct devlink_fmsg *fmsg) +{ + int err; + + err = devlink_fmsg_obj_nest_end(fmsg); + if (err) + return err; + + err = devlink_fmsg_pair_nest_end(fmsg); + if (err) + return err; + + return 0; +} + +int mlx5e_reporter_cq_diagnose(struct mlx5e_cq *cq, struct devlink_fmsg *fmsg) +{ + struct mlx5e_priv *priv = cq->channel->priv; + u32 out[MLX5_ST_SZ_DW(query_cq_out)] = {}; + u8 hw_status; + void *cqc; + int err; + + err = mlx5_core_query_cq(priv->mdev, &cq->mcq, out, sizeof(out)); + if (err) + return err; + + cqc = MLX5_ADDR_OF(query_cq_out, out, cq_context); + hw_status = MLX5_GET(cqc, cqc, status); + + err = mlx5e_reporter_named_obj_nest_start(fmsg, "CQ"); + if (err) + return err; + + err = devlink_fmsg_u32_pair_put(fmsg, "cqn", cq->mcq.cqn); + if (err) + return err; + + err = devlink_fmsg_u8_pair_put(fmsg, "HW status", hw_status); + if (err) + return err; + + err = mlx5e_reporter_named_obj_nest_end(fmsg); + if (err) + return err; + + return 0; +} + +int mlx5e_reporter_cq_common_diagnose(struct mlx5e_cq *cq, struct devlink_fmsg *fmsg) +{ + u8 cq_log_stride; + u32 cq_sz; + int err; + + cq_sz = mlx5_cqwq_get_size(&cq->wq); + cq_log_stride = mlx5_cqwq_get_log_stride_size(&cq->wq); + + err = mlx5e_reporter_named_obj_nest_start(fmsg, "CQ"); + if (err) + return err; + + err = devlink_fmsg_u64_pair_put(fmsg, "stride size", BIT(cq_log_stride)); + if (err) + return err; + + err = devlink_fmsg_u32_pair_put(fmsg, "size", cq_sz); + if (err) + return err; + + err = mlx5e_reporter_named_obj_nest_end(fmsg); + if (err) + return err; + + return 0; +} + +int mlx5e_health_create_reporters(struct mlx5e_priv *priv) +{ + int err; + + err = mlx5e_reporter_tx_create(priv); + if (err) + return err; + + err = mlx5e_reporter_rx_create(priv); + if (err) + return err; + + return 0; +} + +void mlx5e_health_destroy_reporters(struct mlx5e_priv *priv) +{ + mlx5e_reporter_rx_destroy(priv); + mlx5e_reporter_tx_destroy(priv); +} + +void mlx5e_health_channels_update(struct mlx5e_priv *priv) +{ + if (priv->tx_reporter) + devlink_health_reporter_state_update(priv->tx_reporter, + DEVLINK_HEALTH_REPORTER_STATE_HEALTHY); + if (priv->rx_reporter) + devlink_health_reporter_state_update(priv->rx_reporter, + DEVLINK_HEALTH_REPORTER_STATE_HEALTHY); +} + +int mlx5e_health_sq_to_ready(struct mlx5e_channel *channel, u32 sqn) +{ + struct mlx5_core_dev *mdev = channel->mdev; + struct net_device *dev = channel->netdev; + struct mlx5e_modify_sq_param msp = {}; + int err; + + msp.curr_state = MLX5_SQC_STATE_ERR; + msp.next_state = MLX5_SQC_STATE_RST; + + err = mlx5e_modify_sq(mdev, sqn, &msp); + if (err) { + netdev_err(dev, "Failed to move sq 0x%x to reset\n", sqn); + return err; + } + + memset(&msp, 0, sizeof(msp)); + msp.curr_state = MLX5_SQC_STATE_RST; + msp.next_state = MLX5_SQC_STATE_RDY; + + err = mlx5e_modify_sq(mdev, sqn, &msp); + if (err) { + netdev_err(dev, "Failed to move sq 0x%x to ready\n", sqn); + return err; + } + + return 0; +} + +int mlx5e_health_recover_channels(struct mlx5e_priv *priv) +{ + int err = 0; + + rtnl_lock(); + mutex_lock(&priv->state_lock); + + if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) + goto out; + + err = mlx5e_safe_reopen_channels(priv); + +out: + mutex_unlock(&priv->state_lock); + rtnl_unlock(); + + return err; +} + +int mlx5e_health_channel_eq_recover(struct mlx5_eq_comp *eq, struct mlx5e_channel *channel) +{ + u32 eqe_count; + + netdev_err(channel->netdev, "EQ 0x%x: Cons = 0x%x, irqn = 0x%x\n", + eq->core.eqn, eq->core.cons_index, eq->core.irqn); + + eqe_count = mlx5_eq_poll_irq_disabled(eq); + if (!eqe_count) + return -EIO; + + netdev_err(channel->netdev, "Recovered %d eqes on EQ 0x%x\n", + eqe_count, eq->core.eqn); + + channel->stats->eq_rearm++; + return 0; +} + +int mlx5e_health_report(struct mlx5e_priv *priv, + struct devlink_health_reporter *reporter, char *err_str, + struct mlx5e_err_ctx *err_ctx) +{ + if (!reporter) { + netdev_err(priv->netdev, err_str); + return err_ctx->recover(&err_ctx->ctx); + } + return devlink_health_report(reporter, err_str, err_ctx); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/health.h b/drivers/net/ethernet/mellanox/mlx5/core/en/health.h new file mode 100644 index 000000000000..d3693fa547ac --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/health.h @@ -0,0 +1,53 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* Copyright (c) 2019 Mellanox Technologies. */ + +#ifndef __MLX5E_EN_HEALTH_H +#define __MLX5E_EN_HEALTH_H + +#include "en.h" + +#define MLX5E_RX_ERR_CQE(cqe) (get_cqe_opcode(cqe) != MLX5_CQE_RESP_SEND) + +static inline bool cqe_syndrome_needs_recover(u8 syndrome) +{ + return syndrome == MLX5_CQE_SYNDROME_LOCAL_LENGTH_ERR || + syndrome == MLX5_CQE_SYNDROME_LOCAL_QP_OP_ERR || + syndrome == MLX5_CQE_SYNDROME_LOCAL_PROT_ERR || + syndrome == MLX5_CQE_SYNDROME_WR_FLUSH_ERR; +} + +int mlx5e_reporter_tx_create(struct mlx5e_priv *priv); +void mlx5e_reporter_tx_destroy(struct mlx5e_priv *priv); +void mlx5e_reporter_tx_err_cqe(struct mlx5e_txqsq *sq); +int mlx5e_reporter_tx_timeout(struct mlx5e_txqsq *sq); + +int mlx5e_reporter_cq_diagnose(struct mlx5e_cq *cq, struct devlink_fmsg *fmsg); +int mlx5e_reporter_cq_common_diagnose(struct mlx5e_cq *cq, struct devlink_fmsg *fmsg); +int mlx5e_reporter_named_obj_nest_start(struct devlink_fmsg *fmsg, char *name); +int mlx5e_reporter_named_obj_nest_end(struct devlink_fmsg *fmsg); + +int mlx5e_reporter_rx_create(struct mlx5e_priv *priv); +void mlx5e_reporter_rx_destroy(struct mlx5e_priv *priv); +void mlx5e_reporter_icosq_cqe_err(struct mlx5e_icosq *icosq); +void mlx5e_reporter_rq_cqe_err(struct mlx5e_rq *rq); +void mlx5e_reporter_rx_timeout(struct mlx5e_rq *rq); + +#define MLX5E_REPORTER_PER_Q_MAX_LEN 256 + +struct mlx5e_err_ctx { + int (*recover)(void *ctx); + void *ctx; +}; + +int mlx5e_health_sq_to_ready(struct mlx5e_channel *channel, u32 sqn); +int mlx5e_health_channel_eq_recover(struct mlx5_eq_comp *eq, struct mlx5e_channel *channel); +int mlx5e_health_recover_channels(struct mlx5e_priv *priv); +int mlx5e_health_report(struct mlx5e_priv *priv, + struct devlink_health_reporter *reporter, char *err_str, + struct mlx5e_err_ctx *err_ctx); +int mlx5e_health_create_reporters(struct mlx5e_priv *priv); +void mlx5e_health_destroy_reporters(struct mlx5e_priv *priv); +void mlx5e_health_channels_update(struct mlx5e_priv *priv); + + +#endif diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/hv_vhca_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en/hv_vhca_stats.c new file mode 100644 index 000000000000..b3a249b2a482 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/hv_vhca_stats.c @@ -0,0 +1,161 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB +// Copyright (c) 2018 Mellanox Technologies + +#include "en.h" +#include "en/hv_vhca_stats.h" +#include "lib/hv_vhca.h" +#include "lib/hv.h" + +struct mlx5e_hv_vhca_per_ring_stats { + u64 rx_packets; + u64 rx_bytes; + u64 tx_packets; + u64 tx_bytes; +}; + +static void +mlx5e_hv_vhca_fill_ring_stats(struct mlx5e_priv *priv, int ch, + struct mlx5e_hv_vhca_per_ring_stats *data) +{ + struct mlx5e_channel_stats *stats; + int tc; + + stats = &priv->channel_stats[ch]; + data->rx_packets = stats->rq.packets; + data->rx_bytes = stats->rq.bytes; + + for (tc = 0; tc < priv->max_opened_tc; tc++) { + data->tx_packets += stats->sq[tc].packets; + data->tx_bytes += stats->sq[tc].bytes; + } +} + +static void mlx5e_hv_vhca_fill_stats(struct mlx5e_priv *priv, void *data, + int buf_len) +{ + int ch, i = 0; + + for (ch = 0; ch < priv->max_nch; ch++) { + void *buf = data + i; + + if (WARN_ON_ONCE(buf + + sizeof(struct mlx5e_hv_vhca_per_ring_stats) > + data + buf_len)) + return; + + mlx5e_hv_vhca_fill_ring_stats(priv, ch, buf); + i += sizeof(struct mlx5e_hv_vhca_per_ring_stats); + } +} + +static int mlx5e_hv_vhca_stats_buf_size(struct mlx5e_priv *priv) +{ + return (sizeof(struct mlx5e_hv_vhca_per_ring_stats) * + priv->max_nch); +} + +static void mlx5e_hv_vhca_stats_work(struct work_struct *work) +{ + struct mlx5e_hv_vhca_stats_agent *sagent; + struct mlx5_hv_vhca_agent *agent; + struct delayed_work *dwork; + struct mlx5e_priv *priv; + int buf_len, rc; + void *buf; + + dwork = to_delayed_work(work); + sagent = container_of(dwork, struct mlx5e_hv_vhca_stats_agent, work); + priv = container_of(sagent, struct mlx5e_priv, stats_agent); + buf_len = mlx5e_hv_vhca_stats_buf_size(priv); + agent = sagent->agent; + buf = sagent->buf; + + memset(buf, 0, buf_len); + mlx5e_hv_vhca_fill_stats(priv, buf, buf_len); + + rc = mlx5_hv_vhca_agent_write(agent, buf, buf_len); + if (rc) { + mlx5_core_err(priv->mdev, + "%s: Failed to write stats, err = %d\n", + __func__, rc); + return; + } + + if (sagent->delay) + queue_delayed_work(priv->wq, &sagent->work, sagent->delay); +} + +enum { + MLX5_HV_VHCA_STATS_VERSION = 1, + MLX5_HV_VHCA_STATS_UPDATE_ONCE = 0xFFFF, +}; + +static void mlx5e_hv_vhca_stats_control(struct mlx5_hv_vhca_agent *agent, + struct mlx5_hv_vhca_control_block *block) +{ + struct mlx5e_hv_vhca_stats_agent *sagent; + struct mlx5e_priv *priv; + + priv = mlx5_hv_vhca_agent_priv(agent); + sagent = &priv->stats_agent; + + block->version = MLX5_HV_VHCA_STATS_VERSION; + block->rings = priv->max_nch; + + if (!block->command) { + cancel_delayed_work_sync(&priv->stats_agent.work); + return; + } + + sagent->delay = block->command == MLX5_HV_VHCA_STATS_UPDATE_ONCE ? 0 : + msecs_to_jiffies(block->command * 100); + + queue_delayed_work(priv->wq, &sagent->work, sagent->delay); +} + +static void mlx5e_hv_vhca_stats_cleanup(struct mlx5_hv_vhca_agent *agent) +{ + struct mlx5e_priv *priv = mlx5_hv_vhca_agent_priv(agent); + + cancel_delayed_work_sync(&priv->stats_agent.work); +} + +int mlx5e_hv_vhca_stats_create(struct mlx5e_priv *priv) +{ + int buf_len = mlx5e_hv_vhca_stats_buf_size(priv); + struct mlx5_hv_vhca_agent *agent; + + priv->stats_agent.buf = kvzalloc(buf_len, GFP_KERNEL); + if (!priv->stats_agent.buf) + return -ENOMEM; + + agent = mlx5_hv_vhca_agent_create(priv->mdev->hv_vhca, + MLX5_HV_VHCA_AGENT_STATS, + mlx5e_hv_vhca_stats_control, NULL, + mlx5e_hv_vhca_stats_cleanup, + priv); + + if (IS_ERR_OR_NULL(agent)) { + if (IS_ERR(agent)) + netdev_warn(priv->netdev, + "Failed to create hv vhca stats agent, err = %ld\n", + PTR_ERR(agent)); + + kfree(priv->stats_agent.buf); + return IS_ERR_OR_NULL(agent); + } + + priv->stats_agent.agent = agent; + INIT_DELAYED_WORK(&priv->stats_agent.work, mlx5e_hv_vhca_stats_work); + + return 0; +} + +void mlx5e_hv_vhca_stats_destroy(struct mlx5e_priv *priv) +{ + if (IS_ERR_OR_NULL(priv->stats_agent.agent)) + return; + + mlx5_hv_vhca_agent_destroy(priv->stats_agent.agent); + kfree(priv->stats_agent.buf); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/hv_vhca_stats.h b/drivers/net/ethernet/mellanox/mlx5/core/en/hv_vhca_stats.h new file mode 100644 index 000000000000..664463faf77b --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/hv_vhca_stats.h @@ -0,0 +1,25 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* Copyright (c) 2019 Mellanox Technologies. */ + +#ifndef __MLX5_EN_STATS_VHCA_H__ +#define __MLX5_EN_STATS_VHCA_H__ +#include "en.h" + +#if IS_ENABLED(CONFIG_PCI_HYPERV_INTERFACE) + +int mlx5e_hv_vhca_stats_create(struct mlx5e_priv *priv); +void mlx5e_hv_vhca_stats_destroy(struct mlx5e_priv *priv); + +#else + +static inline int mlx5e_hv_vhca_stats_create(struct mlx5e_priv *priv) +{ + return 0; +} + +static inline void mlx5e_hv_vhca_stats_destroy(struct mlx5e_priv *priv) +{ +} +#endif + +#endif /* __MLX5_EN_STATS_VHCA_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/params.c b/drivers/net/ethernet/mellanox/mlx5/core/en/params.c index 79301d116667..eb2e1f2138e4 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/params.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/params.c @@ -25,18 +25,33 @@ u16 mlx5e_get_linear_rq_headroom(struct mlx5e_params *params, return headroom; } -u32 mlx5e_rx_get_linear_frag_sz(struct mlx5e_params *params, - struct mlx5e_xsk_param *xsk) +u32 mlx5e_rx_get_min_frag_sz(struct mlx5e_params *params, + struct mlx5e_xsk_param *xsk) { u32 hw_mtu = MLX5E_SW2HW_MTU(params, params->sw_mtu); u16 linear_rq_headroom = mlx5e_get_linear_rq_headroom(params, xsk); - u32 frag_sz = linear_rq_headroom + hw_mtu; + + return linear_rq_headroom + hw_mtu; +} + +u32 mlx5e_rx_get_linear_frag_sz(struct mlx5e_params *params, + struct mlx5e_xsk_param *xsk) +{ + u32 frag_sz = mlx5e_rx_get_min_frag_sz(params, xsk); /* AF_XDP doesn't build SKBs in place. */ if (!xsk) frag_sz = MLX5_SKB_FRAG_SZ(frag_sz); - /* XDP in mlx5e doesn't support multiple packets per page. */ + /* XDP in mlx5e doesn't support multiple packets per page. AF_XDP is a + * special case. It can run with frames smaller than a page, as it + * doesn't allocate pages dynamically. However, here we pretend that + * fragments are page-sized: it allows to treat XSK frames like pages + * by redirecting alloc and free operations to XSK rings and by using + * the fact there are no multiple packets per "page" (which is a frame). + * The latter is important, because frames may come in a random order, + * and we will have trouble assemblying a real page of multiple frames. + */ if (mlx5e_rx_is_xdp(params, xsk)) frag_sz = max_t(u32, frag_sz, PAGE_SIZE); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/params.h b/drivers/net/ethernet/mellanox/mlx5/core/en/params.h index 3a615d663d84..989d8f429438 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/params.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/params.h @@ -76,6 +76,8 @@ static inline bool mlx5e_qid_validate(const struct mlx5e_profile *profile, u16 mlx5e_get_linear_rq_headroom(struct mlx5e_params *params, struct mlx5e_xsk_param *xsk); +u32 mlx5e_rx_get_min_frag_sz(struct mlx5e_params *params, + struct mlx5e_xsk_param *xsk); u32 mlx5e_rx_get_linear_frag_sz(struct mlx5e_params *params, struct mlx5e_xsk_param *xsk); u8 mlx5e_mpwqe_log_pkts_per_wqe(struct mlx5e_params *params, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter.h b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter.h deleted file mode 100644 index e78e92753d73..000000000000 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter.h +++ /dev/null @@ -1,15 +0,0 @@ -/* SPDX-License-Identifier: GPL-2.0 */ -/* Copyright (c) 2019 Mellanox Technologies. */ - -#ifndef __MLX5E_EN_REPORTER_H -#define __MLX5E_EN_REPORTER_H - -#include <linux/mlx5/driver.h> -#include "en.h" - -int mlx5e_tx_reporter_create(struct mlx5e_priv *priv); -void mlx5e_tx_reporter_destroy(struct mlx5e_priv *priv); -void mlx5e_tx_reporter_err_cqe(struct mlx5e_txqsq *sq); -int mlx5e_tx_reporter_timeout(struct mlx5e_txqsq *sq); - -#endif diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c new file mode 100644 index 000000000000..b860569d4247 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c @@ -0,0 +1,404 @@ +// SPDX-License-Identifier: GPL-2.0 +// Copyright (c) 2019 Mellanox Technologies. + +#include "health.h" +#include "params.h" + +static int mlx5e_query_rq_state(struct mlx5_core_dev *dev, u32 rqn, u8 *state) +{ + int outlen = MLX5_ST_SZ_BYTES(query_rq_out); + void *out; + void *rqc; + int err; + + out = kvzalloc(outlen, GFP_KERNEL); + if (!out) + return -ENOMEM; + + err = mlx5_core_query_rq(dev, rqn, out); + if (err) + goto out; + + rqc = MLX5_ADDR_OF(query_rq_out, out, rq_context); + *state = MLX5_GET(rqc, rqc, state); + +out: + kvfree(out); + return err; +} + +static int mlx5e_wait_for_icosq_flush(struct mlx5e_icosq *icosq) +{ + unsigned long exp_time = jiffies + msecs_to_jiffies(2000); + + while (time_before(jiffies, exp_time)) { + if (icosq->cc == icosq->pc) + return 0; + + msleep(20); + } + + netdev_err(icosq->channel->netdev, + "Wait for ICOSQ 0x%x flush timeout (cc = 0x%x, pc = 0x%x)\n", + icosq->sqn, icosq->cc, icosq->pc); + + return -ETIMEDOUT; +} + +static void mlx5e_reset_icosq_cc_pc(struct mlx5e_icosq *icosq) +{ + WARN_ONCE(icosq->cc != icosq->pc, "ICOSQ 0x%x: cc (0x%x) != pc (0x%x)\n", + icosq->sqn, icosq->cc, icosq->pc); + icosq->cc = 0; + icosq->pc = 0; +} + +static int mlx5e_rx_reporter_err_icosq_cqe_recover(void *ctx) +{ + struct mlx5_core_dev *mdev; + struct mlx5e_icosq *icosq; + struct net_device *dev; + struct mlx5e_rq *rq; + u8 state; + int err; + + icosq = ctx; + rq = &icosq->channel->rq; + mdev = icosq->channel->mdev; + dev = icosq->channel->netdev; + err = mlx5_core_query_sq_state(mdev, icosq->sqn, &state); + if (err) { + netdev_err(dev, "Failed to query ICOSQ 0x%x state. err = %d\n", + icosq->sqn, err); + goto out; + } + + if (state != MLX5_SQC_STATE_ERR) + goto out; + + mlx5e_deactivate_rq(rq); + err = mlx5e_wait_for_icosq_flush(icosq); + if (err) + goto out; + + mlx5e_deactivate_icosq(icosq); + + /* At this point, both the rq and the icosq are disabled */ + + err = mlx5e_health_sq_to_ready(icosq->channel, icosq->sqn); + if (err) + goto out; + + mlx5e_reset_icosq_cc_pc(icosq); + mlx5e_free_rx_descs(rq); + clear_bit(MLX5E_SQ_STATE_RECOVERING, &icosq->state); + mlx5e_activate_icosq(icosq); + mlx5e_activate_rq(rq); + + rq->stats->recover++; + return 0; +out: + clear_bit(MLX5E_SQ_STATE_RECOVERING, &icosq->state); + return err; +} + +void mlx5e_reporter_icosq_cqe_err(struct mlx5e_icosq *icosq) +{ + struct mlx5e_priv *priv = icosq->channel->priv; + char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN]; + struct mlx5e_err_ctx err_ctx = {}; + + err_ctx.ctx = icosq; + err_ctx.recover = mlx5e_rx_reporter_err_icosq_cqe_recover; + sprintf(err_str, "ERR CQE on ICOSQ: 0x%x", icosq->sqn); + + mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx); +} + +static int mlx5e_rq_to_ready(struct mlx5e_rq *rq, int curr_state) +{ + struct net_device *dev = rq->netdev; + int err; + + err = mlx5e_modify_rq_state(rq, curr_state, MLX5_RQC_STATE_RST); + if (err) { + netdev_err(dev, "Failed to move rq 0x%x to reset\n", rq->rqn); + return err; + } + err = mlx5e_modify_rq_state(rq, MLX5_RQC_STATE_RST, MLX5_RQC_STATE_RDY); + if (err) { + netdev_err(dev, "Failed to move rq 0x%x to ready\n", rq->rqn); + return err; + } + + return 0; +} + +static int mlx5e_rx_reporter_err_rq_cqe_recover(void *ctx) +{ + struct mlx5_core_dev *mdev; + struct net_device *dev; + struct mlx5e_rq *rq; + u8 state; + int err; + + rq = ctx; + mdev = rq->mdev; + dev = rq->netdev; + err = mlx5e_query_rq_state(mdev, rq->rqn, &state); + if (err) { + netdev_err(dev, "Failed to query RQ 0x%x state. err = %d\n", + rq->rqn, err); + goto out; + } + + if (state != MLX5_RQC_STATE_ERR) + goto out; + + mlx5e_deactivate_rq(rq); + mlx5e_free_rx_descs(rq); + + err = mlx5e_rq_to_ready(rq, MLX5_RQC_STATE_ERR); + if (err) + goto out; + + clear_bit(MLX5E_RQ_STATE_RECOVERING, &rq->state); + mlx5e_activate_rq(rq); + rq->stats->recover++; + return 0; +out: + clear_bit(MLX5E_RQ_STATE_RECOVERING, &rq->state); + return err; +} + +void mlx5e_reporter_rq_cqe_err(struct mlx5e_rq *rq) +{ + struct mlx5e_priv *priv = rq->channel->priv; + char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN]; + struct mlx5e_err_ctx err_ctx = {}; + + err_ctx.ctx = rq; + err_ctx.recover = mlx5e_rx_reporter_err_rq_cqe_recover; + sprintf(err_str, "ERR CQE on RQ: 0x%x", rq->rqn); + + mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx); +} + +static int mlx5e_rx_reporter_timeout_recover(void *ctx) +{ + struct mlx5e_icosq *icosq; + struct mlx5_eq_comp *eq; + struct mlx5e_rq *rq; + int err; + + rq = ctx; + icosq = &rq->channel->icosq; + eq = rq->cq.mcq.eq; + err = mlx5e_health_channel_eq_recover(eq, rq->channel); + if (err) + clear_bit(MLX5E_SQ_STATE_ENABLED, &icosq->state); + + return err; +} + +void mlx5e_reporter_rx_timeout(struct mlx5e_rq *rq) +{ + struct mlx5e_icosq *icosq = &rq->channel->icosq; + struct mlx5e_priv *priv = rq->channel->priv; + char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN]; + struct mlx5e_err_ctx err_ctx = {}; + + err_ctx.ctx = rq; + err_ctx.recover = mlx5e_rx_reporter_timeout_recover; + sprintf(err_str, "RX timeout on channel: %d, ICOSQ: 0x%x RQ: 0x%x, CQ: 0x%x\n", + icosq->channel->ix, icosq->sqn, rq->rqn, rq->cq.mcq.cqn); + + mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx); +} + +static int mlx5e_rx_reporter_recover_from_ctx(struct mlx5e_err_ctx *err_ctx) +{ + return err_ctx->recover(err_ctx->ctx); +} + +static int mlx5e_rx_reporter_recover(struct devlink_health_reporter *reporter, + void *context) +{ + struct mlx5e_priv *priv = devlink_health_reporter_priv(reporter); + struct mlx5e_err_ctx *err_ctx = context; + + return err_ctx ? mlx5e_rx_reporter_recover_from_ctx(err_ctx) : + mlx5e_health_recover_channels(priv); +} + +static int mlx5e_rx_reporter_build_diagnose_output(struct mlx5e_rq *rq, + struct devlink_fmsg *fmsg) +{ + struct mlx5e_priv *priv = rq->channel->priv; + struct mlx5e_params *params; + struct mlx5e_icosq *icosq; + u8 icosq_hw_state; + int wqes_sz; + u8 hw_state; + u16 wq_head; + int err; + + params = &priv->channels.params; + icosq = &rq->channel->icosq; + err = mlx5e_query_rq_state(priv->mdev, rq->rqn, &hw_state); + if (err) + return err; + + err = mlx5_core_query_sq_state(priv->mdev, icosq->sqn, &icosq_hw_state); + if (err) + return err; + + wqes_sz = mlx5e_rqwq_get_cur_sz(rq); + wq_head = params->rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ ? + rq->mpwqe.wq.head : mlx5_wq_cyc_get_head(&rq->wqe.wq); + + err = devlink_fmsg_obj_nest_start(fmsg); + if (err) + return err; + + err = devlink_fmsg_u32_pair_put(fmsg, "channel ix", rq->channel->ix); + if (err) + return err; + + err = devlink_fmsg_u32_pair_put(fmsg, "rqn", rq->rqn); + if (err) + return err; + + err = devlink_fmsg_u8_pair_put(fmsg, "HW state", hw_state); + if (err) + return err; + + err = devlink_fmsg_u8_pair_put(fmsg, "SW state", rq->state); + if (err) + return err; + + err = devlink_fmsg_u32_pair_put(fmsg, "posted WQEs", wqes_sz); + if (err) + return err; + + err = devlink_fmsg_u32_pair_put(fmsg, "cc", wq_head); + if (err) + return err; + + err = devlink_fmsg_u8_pair_put(fmsg, "ICOSQ HW state", icosq_hw_state); + if (err) + return err; + + err = mlx5e_reporter_cq_diagnose(&rq->cq, fmsg); + if (err) + return err; + + err = devlink_fmsg_obj_nest_end(fmsg); + if (err) + return err; + + return 0; +} + +static int mlx5e_rx_reporter_diagnose(struct devlink_health_reporter *reporter, + struct devlink_fmsg *fmsg) +{ + struct mlx5e_priv *priv = devlink_health_reporter_priv(reporter); + struct mlx5e_params *params = &priv->channels.params; + struct mlx5e_rq *generic_rq; + u32 rq_stride, rq_sz; + int i, err = 0; + + mutex_lock(&priv->state_lock); + + if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) + goto unlock; + + generic_rq = &priv->channels.c[0]->rq; + rq_sz = mlx5e_rqwq_get_size(generic_rq); + rq_stride = BIT(mlx5e_mpwqe_get_log_stride_size(priv->mdev, params, NULL)); + + err = mlx5e_reporter_named_obj_nest_start(fmsg, "Common config"); + if (err) + goto unlock; + + err = mlx5e_reporter_named_obj_nest_start(fmsg, "RQ"); + if (err) + goto unlock; + + err = devlink_fmsg_u8_pair_put(fmsg, "type", params->rq_wq_type); + if (err) + goto unlock; + + err = devlink_fmsg_u64_pair_put(fmsg, "stride size", rq_stride); + if (err) + goto unlock; + + err = devlink_fmsg_u32_pair_put(fmsg, "size", rq_sz); + if (err) + goto unlock; + + err = mlx5e_reporter_named_obj_nest_end(fmsg); + if (err) + goto unlock; + + err = mlx5e_reporter_cq_common_diagnose(&generic_rq->cq, fmsg); + if (err) + goto unlock; + + err = mlx5e_reporter_named_obj_nest_end(fmsg); + if (err) + goto unlock; + + err = devlink_fmsg_arr_pair_nest_start(fmsg, "RQs"); + if (err) + goto unlock; + + for (i = 0; i < priv->channels.num; i++) { + struct mlx5e_rq *rq = &priv->channels.c[i]->rq; + + err = mlx5e_rx_reporter_build_diagnose_output(rq, fmsg); + if (err) + goto unlock; + } + err = devlink_fmsg_arr_pair_nest_end(fmsg); + if (err) + goto unlock; +unlock: + mutex_unlock(&priv->state_lock); + return err; +} + +static const struct devlink_health_reporter_ops mlx5_rx_reporter_ops = { + .name = "rx", + .recover = mlx5e_rx_reporter_recover, + .diagnose = mlx5e_rx_reporter_diagnose, +}; + +#define MLX5E_REPORTER_RX_GRACEFUL_PERIOD 500 + +int mlx5e_reporter_rx_create(struct mlx5e_priv *priv) +{ + struct devlink *devlink = priv_to_devlink(priv->mdev); + struct devlink_health_reporter *reporter; + + reporter = devlink_health_reporter_create(devlink, + &mlx5_rx_reporter_ops, + MLX5E_REPORTER_RX_GRACEFUL_PERIOD, + true, priv); + if (IS_ERR(reporter)) { + netdev_warn(priv->netdev, "Failed to create rx reporter, err = %ld\n", + PTR_ERR(reporter)); + return PTR_ERR(reporter); + } + priv->rx_reporter = reporter; + return 0; +} + +void mlx5e_reporter_rx_destroy(struct mlx5e_priv *priv) +{ + if (!priv->rx_reporter) + return; + + devlink_health_reporter_destroy(priv->rx_reporter); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c index f3d98748b211..bfed558637c2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c @@ -1,16 +1,7 @@ /* SPDX-License-Identifier: GPL-2.0 */ /* Copyright (c) 2019 Mellanox Technologies. */ -#include <net/devlink.h> -#include "reporter.h" -#include "lib/eq.h" - -#define MLX5E_TX_REPORTER_PER_SQ_MAX_LEN 256 - -struct mlx5e_tx_err_ctx { - int (*recover)(struct mlx5e_txqsq *sq); - struct mlx5e_txqsq *sq; -}; +#include "health.h" static int mlx5e_wait_for_sq_flush(struct mlx5e_txqsq *sq) { @@ -40,42 +31,18 @@ static void mlx5e_reset_txqsq_cc_pc(struct mlx5e_txqsq *sq) sq->pc = 0; } -static int mlx5e_sq_to_ready(struct mlx5e_txqsq *sq, int curr_state) +static int mlx5e_tx_reporter_err_cqe_recover(void *ctx) { - struct mlx5_core_dev *mdev = sq->channel->mdev; - struct net_device *dev = sq->channel->netdev; - struct mlx5e_modify_sq_param msp = {0}; - int err; - - msp.curr_state = curr_state; - msp.next_state = MLX5_SQC_STATE_RST; - - err = mlx5e_modify_sq(mdev, sq->sqn, &msp); - if (err) { - netdev_err(dev, "Failed to move sq 0x%x to reset\n", sq->sqn); - return err; - } - - memset(&msp, 0, sizeof(msp)); - msp.curr_state = MLX5_SQC_STATE_RST; - msp.next_state = MLX5_SQC_STATE_RDY; - - err = mlx5e_modify_sq(mdev, sq->sqn, &msp); - if (err) { - netdev_err(dev, "Failed to move sq 0x%x to ready\n", sq->sqn); - return err; - } - - return 0; -} - -static int mlx5e_tx_reporter_err_cqe_recover(struct mlx5e_txqsq *sq) -{ - struct mlx5_core_dev *mdev = sq->channel->mdev; - struct net_device *dev = sq->channel->netdev; + struct mlx5_core_dev *mdev; + struct net_device *dev; + struct mlx5e_txqsq *sq; u8 state; int err; + sq = ctx; + mdev = sq->channel->mdev; + dev = sq->channel->netdev; + if (!test_bit(MLX5E_SQ_STATE_RECOVERING, &sq->state)) return 0; @@ -83,145 +50,130 @@ static int mlx5e_tx_reporter_err_cqe_recover(struct mlx5e_txqsq *sq) if (err) { netdev_err(dev, "Failed to query SQ 0x%x state. err = %d\n", sq->sqn, err); - return err; + goto out; } - if (state != MLX5_SQC_STATE_ERR) { - netdev_err(dev, "SQ 0x%x not in ERROR state\n", sq->sqn); - return -EINVAL; - } + if (state != MLX5_SQC_STATE_ERR) + goto out; mlx5e_tx_disable_queue(sq->txq); err = mlx5e_wait_for_sq_flush(sq); if (err) - return err; + goto out; /* At this point, no new packets will arrive from the stack as TXQ is * marked with QUEUE_STATE_DRV_XOFF. In addition, NAPI cleared all * pending WQEs. SQ can safely reset the SQ. */ - err = mlx5e_sq_to_ready(sq, state); + err = mlx5e_health_sq_to_ready(sq->channel, sq->sqn); if (err) - return err; + goto out; mlx5e_reset_txqsq_cc_pc(sq); sq->stats->recover++; + clear_bit(MLX5E_SQ_STATE_RECOVERING, &sq->state); mlx5e_activate_txqsq(sq); return 0; +out: + clear_bit(MLX5E_SQ_STATE_RECOVERING, &sq->state); + return err; } -static int mlx5_tx_health_report(struct devlink_health_reporter *tx_reporter, - char *err_str, - struct mlx5e_tx_err_ctx *err_ctx) -{ - if (IS_ERR_OR_NULL(tx_reporter)) { - netdev_err(err_ctx->sq->channel->netdev, err_str); - return err_ctx->recover(err_ctx->sq); - } - - return devlink_health_report(tx_reporter, err_str, err_ctx); -} - -void mlx5e_tx_reporter_err_cqe(struct mlx5e_txqsq *sq) +void mlx5e_reporter_tx_err_cqe(struct mlx5e_txqsq *sq) { - char err_str[MLX5E_TX_REPORTER_PER_SQ_MAX_LEN]; - struct mlx5e_tx_err_ctx err_ctx = {0}; + struct mlx5e_priv *priv = sq->channel->priv; + char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN]; + struct mlx5e_err_ctx err_ctx = {0}; - err_ctx.sq = sq; - err_ctx.recover = mlx5e_tx_reporter_err_cqe_recover; + err_ctx.ctx = sq; + err_ctx.recover = mlx5e_tx_reporter_err_cqe_recover; sprintf(err_str, "ERR CQE on SQ: 0x%x", sq->sqn); - mlx5_tx_health_report(sq->channel->priv->tx_reporter, err_str, - &err_ctx); + mlx5e_health_report(priv, priv->tx_reporter, err_str, &err_ctx); } -static int mlx5e_tx_reporter_timeout_recover(struct mlx5e_txqsq *sq) +static int mlx5e_tx_reporter_timeout_recover(void *ctx) { - struct mlx5_eq_comp *eq = sq->cq.mcq.eq; - u32 eqe_count; - - netdev_err(sq->channel->netdev, "EQ 0x%x: Cons = 0x%x, irqn = 0x%x\n", - eq->core.eqn, eq->core.cons_index, eq->core.irqn); + struct mlx5_eq_comp *eq; + struct mlx5e_txqsq *sq; + int err; - eqe_count = mlx5_eq_poll_irq_disabled(eq); - if (!eqe_count) { + sq = ctx; + eq = sq->cq.mcq.eq; + err = mlx5e_health_channel_eq_recover(eq, sq->channel); + if (err) clear_bit(MLX5E_SQ_STATE_ENABLED, &sq->state); - return -EIO; - } - netdev_err(sq->channel->netdev, "Recover %d eqes on EQ 0x%x\n", - eqe_count, eq->core.eqn); - sq->channel->stats->eq_rearm++; - return 0; + return err; } -int mlx5e_tx_reporter_timeout(struct mlx5e_txqsq *sq) +int mlx5e_reporter_tx_timeout(struct mlx5e_txqsq *sq) { - char err_str[MLX5E_TX_REPORTER_PER_SQ_MAX_LEN]; - struct mlx5e_tx_err_ctx err_ctx; + struct mlx5e_priv *priv = sq->channel->priv; + char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN]; + struct mlx5e_err_ctx err_ctx; - err_ctx.sq = sq; - err_ctx.recover = mlx5e_tx_reporter_timeout_recover; + err_ctx.ctx = sq; + err_ctx.recover = mlx5e_tx_reporter_timeout_recover; sprintf(err_str, "TX timeout on queue: %d, SQ: 0x%x, CQ: 0x%x, SQ Cons: 0x%x SQ Prod: 0x%x, usecs since last trans: %u\n", sq->channel->ix, sq->sqn, sq->cq.mcq.cqn, sq->cc, sq->pc, jiffies_to_usecs(jiffies - sq->txq->trans_start)); - return mlx5_tx_health_report(sq->channel->priv->tx_reporter, err_str, - &err_ctx); + return mlx5e_health_report(priv, priv->tx_reporter, err_str, &err_ctx); } /* state lock cannot be grabbed within this function. * It can cause a dead lock or a read-after-free. */ -static int mlx5e_tx_reporter_recover_from_ctx(struct mlx5e_tx_err_ctx *err_ctx) +static int mlx5e_tx_reporter_recover_from_ctx(struct mlx5e_err_ctx *err_ctx) { - return err_ctx->recover(err_ctx->sq); -} - -static int mlx5e_tx_reporter_recover_all(struct mlx5e_priv *priv) -{ - int err = 0; - - rtnl_lock(); - mutex_lock(&priv->state_lock); - - if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) - goto out; - - err = mlx5e_safe_reopen_channels(priv); - -out: - mutex_unlock(&priv->state_lock); - rtnl_unlock(); - - return err; + return err_ctx->recover(err_ctx->ctx); } static int mlx5e_tx_reporter_recover(struct devlink_health_reporter *reporter, void *context) { struct mlx5e_priv *priv = devlink_health_reporter_priv(reporter); - struct mlx5e_tx_err_ctx *err_ctx = context; + struct mlx5e_err_ctx *err_ctx = context; return err_ctx ? mlx5e_tx_reporter_recover_from_ctx(err_ctx) : - mlx5e_tx_reporter_recover_all(priv); + mlx5e_health_recover_channels(priv); } static int mlx5e_tx_reporter_build_diagnose_output(struct devlink_fmsg *fmsg, - u32 sqn, u8 state, bool stopped) + struct mlx5e_txqsq *sq, int tc) { + struct mlx5e_priv *priv = sq->channel->priv; + bool stopped = netif_xmit_stopped(sq->txq); + u8 state; int err; + err = mlx5_core_query_sq_state(priv->mdev, sq->sqn, &state); + if (err) + return err; + err = devlink_fmsg_obj_nest_start(fmsg); if (err) return err; - err = devlink_fmsg_u32_pair_put(fmsg, "sqn", sqn); + err = devlink_fmsg_u32_pair_put(fmsg, "channel ix", sq->ch_ix); + if (err) + return err; + + err = devlink_fmsg_u32_pair_put(fmsg, "tc", tc); + if (err) + return err; + + err = devlink_fmsg_u32_pair_put(fmsg, "txq ix", sq->txq_ix); + if (err) + return err; + + err = devlink_fmsg_u32_pair_put(fmsg, "sqn", sq->sqn); if (err) return err; @@ -233,6 +185,18 @@ mlx5e_tx_reporter_build_diagnose_output(struct devlink_fmsg *fmsg, if (err) return err; + err = devlink_fmsg_u32_pair_put(fmsg, "cc", sq->cc); + if (err) + return err; + + err = devlink_fmsg_u32_pair_put(fmsg, "pc", sq->pc); + if (err) + return err; + + err = mlx5e_reporter_cq_diagnose(&sq->cq, fmsg); + if (err) + return err; + err = devlink_fmsg_obj_nest_end(fmsg); if (err) return err; @@ -244,31 +208,61 @@ static int mlx5e_tx_reporter_diagnose(struct devlink_health_reporter *reporter, struct devlink_fmsg *fmsg) { struct mlx5e_priv *priv = devlink_health_reporter_priv(reporter); - int i, err = 0; + struct mlx5e_txqsq *generic_sq = priv->txq2sq[0]; + u32 sq_stride, sq_sz; + + int i, tc, err = 0; mutex_lock(&priv->state_lock); if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) goto unlock; + sq_sz = mlx5_wq_cyc_get_size(&generic_sq->wq); + sq_stride = MLX5_SEND_WQE_BB; + + err = mlx5e_reporter_named_obj_nest_start(fmsg, "Common Config"); + if (err) + goto unlock; + + err = mlx5e_reporter_named_obj_nest_start(fmsg, "SQ"); + if (err) + goto unlock; + + err = devlink_fmsg_u64_pair_put(fmsg, "stride size", sq_stride); + if (err) + goto unlock; + + err = devlink_fmsg_u32_pair_put(fmsg, "size", sq_sz); + if (err) + goto unlock; + + err = mlx5e_reporter_cq_common_diagnose(&generic_sq->cq, fmsg); + if (err) + goto unlock; + + err = mlx5e_reporter_named_obj_nest_end(fmsg); + if (err) + goto unlock; + + err = mlx5e_reporter_named_obj_nest_end(fmsg); + if (err) + goto unlock; + err = devlink_fmsg_arr_pair_nest_start(fmsg, "SQs"); if (err) goto unlock; - for (i = 0; i < priv->channels.num * priv->channels.params.num_tc; - i++) { - struct mlx5e_txqsq *sq = priv->txq2sq[i]; - u8 state; + for (i = 0; i < priv->channels.num; i++) { + struct mlx5e_channel *c = priv->channels.c[i]; - err = mlx5_core_query_sq_state(priv->mdev, sq->sqn, &state); - if (err) - goto unlock; + for (tc = 0; tc < priv->channels.params.num_tc; tc++) { + struct mlx5e_txqsq *sq = &c->sq[tc]; - err = mlx5e_tx_reporter_build_diagnose_output(fmsg, sq->sqn, - state, - netif_xmit_stopped(sq->txq)); - if (err) - goto unlock; + err = mlx5e_tx_reporter_build_diagnose_output(fmsg, sq, tc); + if (err) + goto unlock; + } } err = devlink_fmsg_arr_pair_nest_end(fmsg); if (err) @@ -287,25 +281,30 @@ static const struct devlink_health_reporter_ops mlx5_tx_reporter_ops = { #define MLX5_REPORTER_TX_GRACEFUL_PERIOD 500 -int mlx5e_tx_reporter_create(struct mlx5e_priv *priv) +int mlx5e_reporter_tx_create(struct mlx5e_priv *priv) { + struct devlink_health_reporter *reporter; struct mlx5_core_dev *mdev = priv->mdev; - struct devlink *devlink = priv_to_devlink(mdev); + struct devlink *devlink; - priv->tx_reporter = + devlink = priv_to_devlink(mdev); + reporter = devlink_health_reporter_create(devlink, &mlx5_tx_reporter_ops, MLX5_REPORTER_TX_GRACEFUL_PERIOD, true, priv); - if (IS_ERR(priv->tx_reporter)) + if (IS_ERR(reporter)) { netdev_warn(priv->netdev, "Failed to create tx reporter, err = %ld\n", - PTR_ERR(priv->tx_reporter)); - return IS_ERR_OR_NULL(priv->tx_reporter); + PTR_ERR(reporter)); + return PTR_ERR(reporter); + } + priv->tx_reporter = reporter; + return 0; } -void mlx5e_tx_reporter_destroy(struct mlx5e_priv *priv) +void mlx5e_reporter_tx_destroy(struct mlx5e_priv *priv) { - if (IS_ERR_OR_NULL(priv->tx_reporter)) + if (!priv->tx_reporter) return; devlink_health_reporter_destroy(priv->tx_reporter); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c index a6a52806be45..f8ee18b4da6f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c @@ -31,29 +31,36 @@ static int get_route_and_out_devs(struct mlx5e_priv *priv, real_dev = is_vlan_dev(dev) ? vlan_dev_real_dev(dev) : dev; uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH); - uplink_upper = netdev_master_upper_dev_get(uplink_dev); + + rcu_read_lock(); + uplink_upper = netdev_master_upper_dev_get_rcu(uplink_dev); + /* mlx5_lag_is_sriov() is a blocking function which can't be called + * while holding rcu read lock. Take the net_device for correctness + * sake. + */ + if (uplink_upper) + dev_hold(uplink_upper); + rcu_read_unlock(); + dst_is_lag_dev = (uplink_upper && netif_is_lag_master(uplink_upper) && real_dev == uplink_upper && mlx5_lag_is_sriov(priv->mdev)); + if (uplink_upper) + dev_put(uplink_upper); /* if the egress device isn't on the same HW e-switch or * it's a LAG device, use the uplink */ + *route_dev = dev; if (!netdev_port_same_parent_id(priv->netdev, real_dev) || - dst_is_lag_dev) { - *route_dev = dev; + dst_is_lag_dev || is_vlan_dev(*route_dev)) *out_dev = uplink_dev; - } else { - *route_dev = dev; - if (is_vlan_dev(*route_dev)) - *out_dev = uplink_dev; - else if (mlx5e_eswitch_rep(dev) && - mlx5e_is_valid_eswitch_fwd_dev(priv, dev)) - *out_dev = *route_dev; - else - return -EOPNOTSUPP; - } + else if (mlx5e_eswitch_rep(dev) && + mlx5e_is_valid_eswitch_fwd_dev(priv, dev)) + *out_dev = *route_dev; + else + return -EOPNOTSUPP; if (!(mlx5e_eswitch_rep(*out_dev) && mlx5e_is_uplink_rep(netdev_priv(*out_dev)))) @@ -284,14 +291,14 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv, */ goto out; } - - err = mlx5_packet_reformat_alloc(priv->mdev, - e->reformat_type, - ipv4_encap_size, encap_header, - MLX5_FLOW_NAMESPACE_FDB, - &e->encap_id); - if (err) + e->pkt_reformat = mlx5_packet_reformat_alloc(priv->mdev, + e->reformat_type, + ipv4_encap_size, encap_header, + MLX5_FLOW_NAMESPACE_FDB); + if (IS_ERR(e->pkt_reformat)) { + err = PTR_ERR(e->pkt_reformat); goto destroy_neigh_entry; + } e->flags |= MLX5_ENCAP_ENTRY_VALID; mlx5e_rep_queue_neigh_stats_work(netdev_priv(out_dev)); @@ -400,13 +407,14 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv, goto out; } - err = mlx5_packet_reformat_alloc(priv->mdev, - e->reformat_type, - ipv6_encap_size, encap_header, - MLX5_FLOW_NAMESPACE_FDB, - &e->encap_id); - if (err) + e->pkt_reformat = mlx5_packet_reformat_alloc(priv->mdev, + e->reformat_type, + ipv6_encap_size, encap_header, + MLX5_FLOW_NAMESPACE_FDB); + if (IS_ERR(e->pkt_reformat)) { + err = PTR_ERR(e->pkt_reformat); goto destroy_neigh_entry; + } e->flags |= MLX5_ENCAP_ENTRY_VALID; mlx5e_rep_queue_neigh_stats_work(netdev_priv(out_dev)); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h b/drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h index ddfe19adb3d9..87be96747902 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h @@ -6,7 +6,7 @@ #include "en.h" -#define MLX5E_SQ_NOPS_ROOM MLX5_SEND_WQE_MAX_WQEBBS +#define MLX5E_SQ_NOPS_ROOM (MLX5_SEND_WQE_MAX_WQEBBS - 1) #define MLX5E_SQ_STOP_ROOM (MLX5_SEND_WQE_MAX_WQEBBS +\ MLX5E_SQ_NOPS_ROOM) @@ -117,9 +117,27 @@ mlx5e_notify_hw(struct mlx5_wq_cyc *wq, u16 pc, void __iomem *uar_map, mlx5_write64((__be32 *)ctrl, uar_map); } -static inline bool mlx5e_transport_inline_tx_wqe(struct mlx5e_tx_wqe *wqe) +static inline bool mlx5e_transport_inline_tx_wqe(struct mlx5_wqe_ctrl_seg *cseg) { - return !!wqe->ctrl.tisn; + return cseg && !!cseg->tisn; +} + +static inline u8 +mlx5e_tx_wqe_inline_mode(struct mlx5e_txqsq *sq, struct mlx5_wqe_ctrl_seg *cseg, + struct sk_buff *skb) +{ + u8 mode; + + if (mlx5e_transport_inline_tx_wqe(cseg)) + return MLX5_INLINE_MODE_TCP_UDP; + + mode = sq->min_inline_mode; + + if (skb_vlan_tag_present(skb) && + test_bit(MLX5E_SQ_STATE_VLAN_NEED_L2_INLINE, &sq->state)) + mode = max_t(u8, MLX5_INLINE_MODE_L2, mode); + + return mode; } static inline void mlx5e_cq_arm(struct mlx5e_cq *cq) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c index b0b982cf69bb..f049e0ac308a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c @@ -122,6 +122,7 @@ bool mlx5e_xdp_handle(struct mlx5e_rq *rq, struct mlx5e_dma_info *di, void *va, u16 *rx_headroom, u32 *len, bool xsk) { struct bpf_prog *prog = READ_ONCE(rq->xdp_prog); + struct xdp_umem *umem = rq->umem; struct xdp_buff xdp; u32 act; int err; @@ -138,8 +139,11 @@ bool mlx5e_xdp_handle(struct mlx5e_rq *rq, struct mlx5e_dma_info *di, xdp.rxq = &rq->xdp_rxq; act = bpf_prog_run_xdp(prog, &xdp); - if (xsk) - xdp.handle += xdp.data - xdp.data_hard_start; + if (xsk) { + u64 off = xdp.data - xdp.data_hard_start; + + xdp.handle = xsk_umem_adjust_offset(umem, xdp.handle, off); + } switch (act) { case XDP_PASS: *rx_headroom = xdp.data - xdp.data_hard_start; @@ -179,33 +183,19 @@ static void mlx5e_xdp_mpwqe_session_start(struct mlx5e_xdpsq *sq) struct mlx5e_xdp_mpwqe *session = &sq->mpwqe; struct mlx5e_xdpsq_stats *stats = sq->stats; struct mlx5_wq_cyc *wq = &sq->wq; - u8 wqebbs; - u16 pi; - - mlx5e_xdpsq_fetch_wqe(sq, &session->wqe); - - prefetchw(session->wqe->data); - session->ds_count = MLX5E_XDP_TX_EMPTY_DS_COUNT; - session->pkt_count = 0; - session->complete = 0; + u16 pi, contig_wqebbs; pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc); + contig_wqebbs = mlx5_wq_cyc_get_contig_wqebbs(wq, pi); -/* The mult of MLX5_SEND_WQE_MAX_WQEBBS * MLX5_SEND_WQEBB_NUM_DS - * (16 * 4 == 64) does not fit in the 6-bit DS field of Ctrl Segment. - * We use a bound lower that MLX5_SEND_WQE_MAX_WQEBBS to let a - * full-session WQE be cache-aligned. - */ -#if L1_CACHE_BYTES < 128 -#define MLX5E_XDP_MPW_MAX_WQEBBS (MLX5_SEND_WQE_MAX_WQEBBS - 1) -#else -#define MLX5E_XDP_MPW_MAX_WQEBBS (MLX5_SEND_WQE_MAX_WQEBBS - 2) -#endif + if (unlikely(contig_wqebbs < MLX5_SEND_WQE_MAX_WQEBBS)) + mlx5e_fill_xdpsq_frag_edge(sq, wq, pi, contig_wqebbs); - wqebbs = min_t(u16, mlx5_wq_cyc_get_contig_wqebbs(wq, pi), - MLX5E_XDP_MPW_MAX_WQEBBS); + session->wqe = mlx5e_xdpsq_fetch_wqe(sq, &pi); - session->max_ds_count = MLX5_SEND_WQEBB_NUM_DS * wqebbs; + prefetchw(session->wqe->data); + session->ds_count = MLX5E_XDP_TX_EMPTY_DS_COUNT; + session->pkt_count = 0; mlx5e_xdp_update_inline_state(sq); @@ -244,7 +234,7 @@ static int mlx5e_xmit_xdp_frame_check_mpwqe(struct mlx5e_xdpsq *sq) { if (unlikely(!sq->mpwqe.wqe)) { if (unlikely(!mlx5e_wqc_has_room_for(&sq->wq, sq->cc, sq->pc, - MLX5_SEND_WQE_MAX_WQEBBS))) { + MLX5E_XDPSQ_STOP_ROOM))) { /* SQ is full, ring doorbell */ mlx5e_xmit_xdp_doorbell(sq); sq->stats->full++; @@ -285,8 +275,8 @@ static bool mlx5e_xmit_xdp_frame_mpwqe(struct mlx5e_xdpsq *sq, mlx5e_xdp_mpwqe_add_dseg(sq, xdptxd, stats); - if (unlikely(session->complete || - session->ds_count == session->max_ds_count)) + if (unlikely(mlx5e_xdp_no_room_for_inline_pkt(session) || + session->ds_count == MLX5E_XDP_MPW_MAX_NUM_DS)) mlx5e_xdp_mpwqe_complete(sq); mlx5e_xdpi_fifo_push(&sq->db.xdpi_fifo, xdpi); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h index b90923932668..36ac1e3816b9 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h @@ -40,6 +40,26 @@ (sizeof(struct mlx5e_tx_wqe) / MLX5_SEND_WQE_DS) #define MLX5E_XDP_TX_DS_COUNT (MLX5E_XDP_TX_EMPTY_DS_COUNT + 1 /* SG DS */) +#define MLX5E_XDPSQ_STOP_ROOM (MLX5E_SQ_STOP_ROOM) + +#define MLX5E_XDP_INLINE_WQE_SZ_THRSD (256 - sizeof(struct mlx5_wqe_inline_seg)) +#define MLX5E_XDP_INLINE_WQE_MAX_DS_CNT \ + DIV_ROUND_UP(MLX5E_XDP_INLINE_WQE_SZ_THRSD, MLX5_SEND_WQE_DS) + +/* The mult of MLX5_SEND_WQE_MAX_WQEBBS * MLX5_SEND_WQEBB_NUM_DS + * (16 * 4 == 64) does not fit in the 6-bit DS field of Ctrl Segment. + * We use a bound lower that MLX5_SEND_WQE_MAX_WQEBBS to let a + * full-session WQE be cache-aligned. + */ +#if L1_CACHE_BYTES < 128 +#define MLX5E_XDP_MPW_MAX_WQEBBS (MLX5_SEND_WQE_MAX_WQEBBS - 1) +#else +#define MLX5E_XDP_MPW_MAX_WQEBBS (MLX5_SEND_WQE_MAX_WQEBBS - 2) +#endif + +#define MLX5E_XDP_MPW_MAX_NUM_DS \ + (MLX5E_XDP_MPW_MAX_WQEBBS * MLX5_SEND_WQEBB_NUM_DS) + struct mlx5e_xsk_param; int mlx5e_xdp_max_mtu(struct mlx5e_params *params, struct mlx5e_xsk_param *xsk); bool mlx5e_xdp_handle(struct mlx5e_rq *rq, struct mlx5e_dma_info *di, @@ -114,6 +134,30 @@ static inline void mlx5e_xdp_update_inline_state(struct mlx5e_xdpsq *sq) session->inline_on = 1; } +static inline bool +mlx5e_xdp_no_room_for_inline_pkt(struct mlx5e_xdp_mpwqe *session) +{ + return session->inline_on && + session->ds_count + MLX5E_XDP_INLINE_WQE_MAX_DS_CNT > MLX5E_XDP_MPW_MAX_NUM_DS; +} + +static inline void +mlx5e_fill_xdpsq_frag_edge(struct mlx5e_xdpsq *sq, struct mlx5_wq_cyc *wq, + u16 pi, u16 nnops) +{ + struct mlx5e_xdp_wqe_info *edge_wi, *wi = &sq->db.wqe_info[pi]; + + edge_wi = wi + nnops; + /* fill sq frag edge with nops to avoid wqe wrapping two pages */ + for (; wi < edge_wi; wi++) { + wi->num_wqebbs = 1; + wi->num_pkts = 0; + mlx5e_post_nop(wq, sq->sqn, &sq->pc); + } + + sq->stats->nops += nnops; +} + static inline void mlx5e_xdp_mpwqe_add_dseg(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_xmit_data *xdptxd, @@ -126,20 +170,12 @@ mlx5e_xdp_mpwqe_add_dseg(struct mlx5e_xdpsq *sq, session->pkt_count++; -#define MLX5E_XDP_INLINE_WQE_SZ_THRSD (256 - sizeof(struct mlx5_wqe_inline_seg)) - if (session->inline_on && dma_len <= MLX5E_XDP_INLINE_WQE_SZ_THRSD) { struct mlx5_wqe_inline_seg *inline_dseg = (struct mlx5_wqe_inline_seg *)dseg; u16 ds_len = sizeof(*inline_dseg) + dma_len; u16 ds_cnt = DIV_ROUND_UP(ds_len, MLX5_SEND_WQE_DS); - if (unlikely(session->ds_count + ds_cnt > session->max_ds_count)) { - /* Not enough space for inline wqe, send with memory pointer */ - session->complete = true; - goto no_inline; - } - inline_dseg->byte_count = cpu_to_be32(dma_len | MLX5_INLINE_SEG); memcpy(inline_dseg->data, xdptxd->data, dma_len); @@ -148,21 +184,23 @@ mlx5e_xdp_mpwqe_add_dseg(struct mlx5e_xdpsq *sq, return; } -no_inline: dseg->addr = cpu_to_be64(xdptxd->dma_addr); dseg->byte_count = cpu_to_be32(dma_len); dseg->lkey = sq->mkey_be; session->ds_count++; } -static inline void mlx5e_xdpsq_fetch_wqe(struct mlx5e_xdpsq *sq, - struct mlx5e_tx_wqe **wqe) +static inline struct mlx5e_tx_wqe * +mlx5e_xdpsq_fetch_wqe(struct mlx5e_xdpsq *sq, u16 *pi) { struct mlx5_wq_cyc *wq = &sq->wq; - u16 pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc); + struct mlx5e_tx_wqe *wqe; + + *pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc); + wqe = mlx5_wq_cyc_get_wqe(wq, *pi); + memset(wqe, 0, sizeof(*wqe)); - *wqe = mlx5_wq_cyc_get_wqe(wq, pi); - memset(*wqe, 0, sizeof(**wqe)); + return wqe; } static inline void diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/Makefile deleted file mode 100644 index 5ee42991900a..000000000000 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/Makefile +++ /dev/null @@ -1 +0,0 @@ -subdir-ccflags-y += -I$(src)/../.. diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.c index 6a55573ec8f2..475b6bd5d29b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.c @@ -24,7 +24,8 @@ int mlx5e_xsk_page_alloc_umem(struct mlx5e_rq *rq, if (!xsk_umem_peek_addr_rq(umem, &handle)) return -ENOMEM; - dma_info->xsk.handle = handle + rq->buff.umem_headroom; + dma_info->xsk.handle = xsk_umem_adjust_offset(umem, handle, + rq->buff.umem_headroom); dma_info->xsk.data = xdp_umem_get_data(umem, dma_info->xsk.handle); /* No need to add headroom to the DMA address. In striding RQ case, we @@ -104,7 +105,7 @@ struct sk_buff *mlx5e_xsk_skb_from_cqe_mpwrq_linear(struct mlx5e_rq *rq, /* head_offset is not used in this function, because di->xsk.data and * di->addr point directly to the necessary place. Furthermore, in the - * current implementation, one page = one packet = one frame, so + * current implementation, UMR pages are mapped to XSK frames, so * head_offset should always be 0. */ WARN_ON_ONCE(head_offset); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.h b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.h index 307b923a1361..cab0e93497ae 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.h @@ -5,6 +5,7 @@ #define __MLX5_EN_XSK_RX_H__ #include "en.h" +#include <net/xdp_sock.h> /* RX data path */ @@ -24,4 +25,17 @@ struct sk_buff *mlx5e_xsk_skb_from_cqe_linear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt); +static inline bool mlx5e_xsk_update_rx_wakeup(struct mlx5e_rq *rq, bool alloc_err) +{ + if (!xsk_umem_uses_need_wakeup(rq->umem)) + return alloc_err; + + if (unlikely(alloc_err)) + xsk_set_rx_need_wakeup(rq->umem); + else + xsk_clear_rx_need_wakeup(rq->umem); + + return false; +} + #endif /* __MLX5_EN_XSK_RX_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/setup.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/setup.c index aaffa6f68dc0..631af8dee517 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/setup.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/setup.c @@ -4,18 +4,23 @@ #include "setup.h" #include "en/params.h" +/* It matches XDP_UMEM_MIN_CHUNK_SIZE, but as this constant is private and may + * change unexpectedly, and mlx5e has a minimum valid stride size for striding + * RQ, keep this check in the driver. + */ +#define MLX5E_MIN_XSK_CHUNK_SIZE 2048 + bool mlx5e_validate_xsk_param(struct mlx5e_params *params, struct mlx5e_xsk_param *xsk, struct mlx5_core_dev *mdev) { - /* AF_XDP doesn't support frames larger than PAGE_SIZE, and the current - * mlx5e XDP implementation doesn't support multiple packets per page. - */ - if (xsk->chunk_size != PAGE_SIZE) + /* AF_XDP doesn't support frames larger than PAGE_SIZE. */ + if (xsk->chunk_size > PAGE_SIZE || + xsk->chunk_size < MLX5E_MIN_XSK_CHUNK_SIZE) return false; /* Current MTU and XSK headroom don't allow packets to fit the frames. */ - if (mlx5e_rx_get_linear_frag_sz(params, xsk) > xsk->chunk_size) + if (mlx5e_rx_get_min_frag_sz(params, xsk) > xsk->chunk_size) return false; /* frag_sz is different for regular and XSK RQs, so ensure that linear @@ -60,24 +65,28 @@ int mlx5e_open_xsk(struct mlx5e_priv *priv, struct mlx5e_params *params, struct mlx5e_xsk_param *xsk, struct xdp_umem *umem, struct mlx5e_channel *c) { - struct mlx5e_channel_param cparam = {}; + struct mlx5e_channel_param *cparam; struct dim_cq_moder icocq_moder = {}; int err; if (!mlx5e_validate_xsk_param(params, xsk, priv->mdev)) return -EINVAL; - mlx5e_build_xsk_cparam(priv, params, xsk, &cparam); + cparam = kvzalloc(sizeof(*cparam), GFP_KERNEL); + if (!cparam) + return -ENOMEM; + + mlx5e_build_xsk_cparam(priv, params, xsk, cparam); - err = mlx5e_open_cq(c, params->rx_cq_moderation, &cparam.rx_cq, &c->xskrq.cq); + err = mlx5e_open_cq(c, params->rx_cq_moderation, &cparam->rx_cq, &c->xskrq.cq); if (unlikely(err)) - return err; + goto err_free_cparam; - err = mlx5e_open_rq(c, params, &cparam.rq, xsk, umem, &c->xskrq); + err = mlx5e_open_rq(c, params, &cparam->rq, xsk, umem, &c->xskrq); if (unlikely(err)) goto err_close_rx_cq; - err = mlx5e_open_cq(c, params->tx_cq_moderation, &cparam.tx_cq, &c->xsksq.cq); + err = mlx5e_open_cq(c, params->tx_cq_moderation, &cparam->tx_cq, &c->xsksq.cq); if (unlikely(err)) goto err_close_rq; @@ -87,21 +96,23 @@ int mlx5e_open_xsk(struct mlx5e_priv *priv, struct mlx5e_params *params, * is disabled and then reenabled, but the SQ continues receiving CQEs * from the old UMEM. */ - err = mlx5e_open_xdpsq(c, params, &cparam.xdp_sq, umem, &c->xsksq, true); + err = mlx5e_open_xdpsq(c, params, &cparam->xdp_sq, umem, &c->xsksq, true); if (unlikely(err)) goto err_close_tx_cq; - err = mlx5e_open_cq(c, icocq_moder, &cparam.icosq_cq, &c->xskicosq.cq); + err = mlx5e_open_cq(c, icocq_moder, &cparam->icosq_cq, &c->xskicosq.cq); if (unlikely(err)) goto err_close_sq; /* Create a dedicated SQ for posting NOPs whenever we need an IRQ to be * triggered and NAPI to be called on the correct CPU. */ - err = mlx5e_open_icosq(c, params, &cparam.icosq, &c->xskicosq); + err = mlx5e_open_icosq(c, params, &cparam->icosq, &c->xskicosq); if (unlikely(err)) goto err_close_icocq; + kvfree(cparam); + spin_lock_init(&c->xskicosq_lock); set_bit(MLX5E_CHANNEL_STATE_XSK, c->state); @@ -123,6 +134,9 @@ err_close_rq: err_close_rx_cq: mlx5e_close_cq(&c->xskrq.cq); +err_free_cparam: + kvfree(cparam); + return err; } @@ -141,15 +155,20 @@ void mlx5e_close_xsk(struct mlx5e_channel *c) void mlx5e_activate_xsk(struct mlx5e_channel *c) { + mlx5e_activate_icosq(&c->xskicosq); set_bit(MLX5E_RQ_STATE_ENABLED, &c->xskrq.state); /* TX queue is created active. */ + + spin_lock(&c->xskicosq_lock); mlx5e_trigger_irq(&c->xskicosq); + spin_unlock(&c->xskicosq_lock); } void mlx5e_deactivate_xsk(struct mlx5e_channel *c) { mlx5e_deactivate_rq(&c->xskrq); /* TX queue is disabled on close. */ + mlx5e_deactivate_icosq(&c->xskicosq); } static int mlx5e_redirect_xsk_rqt(struct mlx5e_priv *priv, u16 ix, u32 rqn) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.c index 35e188cf4ea4..87827477d38c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.c @@ -7,7 +7,7 @@ #include "en/params.h" #include <net/xdp_sock.h> -int mlx5e_xsk_async_xmit(struct net_device *dev, u32 qid) +int mlx5e_xsk_wakeup(struct net_device *dev, u32 qid, u32 flags) { struct mlx5e_priv *priv = netdev_priv(dev); struct mlx5e_params *params = &priv->channels.params; @@ -26,6 +26,13 @@ int mlx5e_xsk_async_xmit(struct net_device *dev, u32 qid) return -ENXIO; if (!napi_if_scheduled_mark_missed(&c->napi)) { + /* To avoid WQE overrun, don't post a NOP if XSKICOSQ is not + * active and not polled by NAPI. Return 0, because the upcoming + * activate will trigger the IRQ for us. + */ + if (unlikely(!test_bit(MLX5E_SQ_STATE_ENABLED, &c->xskicosq.state))) + return 0; + spin_lock(&c->xskicosq_lock); mlx5e_trigger_irq(&c->xskicosq); spin_unlock(&c->xskicosq_lock); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.h b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.h index 7add18bf78d8..79b487d89757 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.h @@ -5,11 +5,23 @@ #define __MLX5_EN_XSK_TX_H__ #include "en.h" +#include <net/xdp_sock.h> /* TX data path */ -int mlx5e_xsk_async_xmit(struct net_device *dev, u32 qid); +int mlx5e_xsk_wakeup(struct net_device *dev, u32 qid, u32 flags); bool mlx5e_xsk_tx(struct mlx5e_xdpsq *sq, unsigned int budget); +static inline void mlx5e_xsk_update_tx_wakeup(struct mlx5e_xdpsq *sq) +{ + if (!xsk_umem_uses_need_wakeup(sq->umem)) + return; + + if (sq->pc != sq->cc) + xsk_clear_tx_need_wakeup(sq->umem); + else + xsk_set_tx_need_wakeup(sq->umem); +} + #endif /* __MLX5_EN_XSK_TX_H__ */ |