From 698114968a22f6c0c9f42e983ba033cc36bb7217 Mon Sep 17 00:00:00 2001 From: Moni Shoua Date: Thu, 8 Nov 2018 21:10:08 +0200 Subject: net/mlx5: Release resource on error flow Fix reference counting leakage when the event handler aborts due to an unsupported event for the resource type. Fixes: a14c2d4beee5 ("net/mlx5_core: Warn on unsupported events of QP/RQ/SQ") Signed-off-by: Moni Shoua Reviewed-by: Majd Dibbiny Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/qp.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/qp.c b/drivers/net/ethernet/mellanox/mlx5/core/qp.c index 91b8139a388d..690dc1dd9391 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/qp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/qp.c @@ -132,7 +132,7 @@ void mlx5_rsc_event(struct mlx5_core_dev *dev, u32 rsn, int event_type) if (!is_event_type_allowed((rsn >> MLX5_USER_INDEX_LEN), event_type)) { mlx5_core_warn(dev, "event 0x%.2x is not allowed on resource 0x%.8x\n", event_type, rsn); - return; + goto out; } switch (common->res) { @@ -150,7 +150,7 @@ void mlx5_rsc_event(struct mlx5_core_dev *dev, u32 rsn, int event_type) default: mlx5_core_warn(dev, "invalid resource type for 0x%x\n", rsn); } - +out: mlx5_core_put_rsc(common); } -- cgit v1.2.3-55-g7522 From 27e95603f4dfec470c6d26bea5174aa71b30e971 Mon Sep 17 00:00:00 2001 From: Moni Shoua Date: Thu, 8 Nov 2018 21:10:10 +0200 Subject: net/mlx5: Add interface to hold and release core resources Sometimes upper layers may want to prevent the destruction of a core resource for a period of time while work on that resource is in progress. Add API to support this. Signed-off-by: Moni Shoua Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/qp.c | 16 ++++++++++++++++ include/linux/mlx5/qp.h | 5 +++++ 2 files changed, 21 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/qp.c b/drivers/net/ethernet/mellanox/mlx5/core/qp.c index 690dc1dd9391..cba4a435043a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/qp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/qp.c @@ -670,3 +670,19 @@ int mlx5_core_query_q_counter(struct mlx5_core_dev *dev, u16 counter_id, return mlx5_cmd_exec(dev, in, sizeof(in), out, out_size); } EXPORT_SYMBOL_GPL(mlx5_core_query_q_counter); + +struct mlx5_core_rsc_common *mlx5_core_res_hold(struct mlx5_core_dev *dev, + int res_num, + enum mlx5_res_type res_type) +{ + u32 rsn = res_num | (res_type << MLX5_USER_INDEX_LEN); + + return mlx5_get_rsc(dev, rsn); +} +EXPORT_SYMBOL_GPL(mlx5_core_res_hold); + +void mlx5_core_res_put(struct mlx5_core_rsc_common *res) +{ + mlx5_core_put_rsc(res); +} +EXPORT_SYMBOL_GPL(mlx5_core_res_put); diff --git a/include/linux/mlx5/qp.h b/include/linux/mlx5/qp.h index fbe322c966bc..b26ea9077384 100644 --- a/include/linux/mlx5/qp.h +++ b/include/linux/mlx5/qp.h @@ -596,6 +596,11 @@ int mlx5_core_dealloc_q_counter(struct mlx5_core_dev *dev, u16 counter_id); int mlx5_core_query_q_counter(struct mlx5_core_dev *dev, u16 counter_id, int reset, void *out, int out_size); +struct mlx5_core_rsc_common *mlx5_core_res_hold(struct mlx5_core_dev *dev, + int res_num, + enum mlx5_res_type res_type); +void mlx5_core_res_put(struct mlx5_core_rsc_common *res); + static inline const char *mlx5_qp_type_str(int type) { switch (type) { -- cgit v1.2.3-55-g7522 From ef90c5e9757d5abc80b947cbcaacc43d1e916488 Mon Sep 17 00:00:00 2001 From: Moni Shoua Date: Thu, 8 Nov 2018 21:10:13 +0200 Subject: net/mlx5: Return success for PAGE_FAULT_RESUME in internal error state When the device is in internal error state, command interface isn't accessible and the driver decides which commands to fail and which to pass. Move the PAGE_FAULT_RESUME command to the pass list in order to avoid redundant failure messages. Fixes: 89d44f0a6c73 ("net/mlx5_core: Add pci error handlers to mlx5_core driver") Signed-off-by: Moni Shoua Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/cmd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c index a5a0823e5ada..7b18aff955f1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c @@ -313,6 +313,7 @@ static int mlx5_internal_err_ret_value(struct mlx5_core_dev *dev, u16 op, case MLX5_CMD_OP_FPGA_DESTROY_QP: case MLX5_CMD_OP_DESTROY_GENERAL_OBJECT: case MLX5_CMD_OP_DEALLOC_MEMIC: + case MLX5_CMD_OP_PAGE_FAULT_RESUME: return MLX5_CMD_STAT_OK; case MLX5_CMD_OP_QUERY_HCA_CAP: @@ -326,7 +327,6 @@ static int mlx5_internal_err_ret_value(struct mlx5_core_dev *dev, u16 op, case MLX5_CMD_OP_CREATE_MKEY: case MLX5_CMD_OP_QUERY_MKEY: case MLX5_CMD_OP_QUERY_SPECIAL_CONTEXTS: - case MLX5_CMD_OP_PAGE_FAULT_RESUME: case MLX5_CMD_OP_CREATE_EQ: case MLX5_CMD_OP_QUERY_EQ: case MLX5_CMD_OP_GEN_EQE: -- cgit v1.2.3-55-g7522 From 90290db7669ba680b37b7006cbf6e5cee6cba779 Mon Sep 17 00:00:00 2001 From: Moni Shoua Date: Thu, 8 Nov 2018 21:10:14 +0200 Subject: net/mlx5: Use multi threaded workqueue for page fault handling Page fault events are processed in a workqueue context. Since each QP can have up to two concurrent unrelated page-faults, one for requester and one for responder, page-fault handling can be done in parallel. Achieve this by changing the workqueue to be multi-threaded. The number of threads is the same as the number of command interface channels to avoid command interface bottlenecks. In addition to multi-threads, change the workqueue flags to give it high priority. Stress benchmark shows that before this change 85% of page faults were waiting in queue 8 seconds or more while after the change 98% of page faults were waiting in queue 64 milliseconds or less. The number of threads was chosen as the number of channels to the command interface. Fixes: d9aaed838765 ("{net,IB}/mlx5: Refactor page fault handling") Signed-off-by: Moni Shoua Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index c1e1a16a9b07..aeab0c4f60f4 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -356,8 +356,9 @@ static int init_pf_ctx(struct mlx5_eq_pagefault *pf_ctx, const char *name) spin_lock_init(&pf_ctx->lock); INIT_WORK(&pf_ctx->work, eq_pf_action); - pf_ctx->wq = alloc_ordered_workqueue(name, - WQ_MEM_RECLAIM); + pf_ctx->wq = alloc_workqueue(name, + WQ_HIGHPRI | WQ_UNBOUND | WQ_MEM_RECLAIM, + MLX5_NUM_CMD_EQE); if (!pf_ctx->wq) return -ENOMEM; -- cgit v1.2.3-55-g7522 From 1e86ace4c140fd5a693e266c9b23409358f25381 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 19 Nov 2018 10:52:31 -0800 Subject: net/mlx5: EQ, Use the right place to store/read IRQ affinity hint Currently the cpu affinity hint mask for completion EQs is stored and read from the wrong place, since reading and storing is done from the same index, there is no actual issue with that, but internal irq_info for completion EQs stars at MLX5_EQ_VEC_COMP_BASE offset in irq_info array, this patch changes the code to use the correct offset to store and read the IRQ affinity hint. Signed-off-by: Saeed Mahameed Reviewed-by: Leon Romanovsky Reviewed-by: Tariq Toukan Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/main.c | 14 ++++++++------ include/linux/mlx5/driver.h | 2 +- 3 files changed, 10 insertions(+), 8 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 1243edbedc9e..2839c30dd3a0 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -1760,7 +1760,7 @@ static void mlx5e_close_cq(struct mlx5e_cq *cq) static int mlx5e_get_cpu(struct mlx5e_priv *priv, int ix) { - return cpumask_first(priv->mdev->priv.irq_info[ix].mask); + return cpumask_first(priv->mdev->priv.irq_info[ix + MLX5_EQ_VEC_COMP_BASE].mask); } static int mlx5e_open_tx_cqs(struct mlx5e_channel *c, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 28132c7dc05f..d5cea0a36e6a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -640,18 +640,19 @@ u64 mlx5_read_internal_timer(struct mlx5_core_dev *dev) static int mlx5_irq_set_affinity_hint(struct mlx5_core_dev *mdev, int i) { struct mlx5_priv *priv = &mdev->priv; - int irq = pci_irq_vector(mdev->pdev, MLX5_EQ_VEC_COMP_BASE + i); + int vecidx = MLX5_EQ_VEC_COMP_BASE + i; + int irq = pci_irq_vector(mdev->pdev, vecidx); - if (!zalloc_cpumask_var(&priv->irq_info[i].mask, GFP_KERNEL)) { + if (!zalloc_cpumask_var(&priv->irq_info[vecidx].mask, GFP_KERNEL)) { mlx5_core_warn(mdev, "zalloc_cpumask_var failed"); return -ENOMEM; } cpumask_set_cpu(cpumask_local_spread(i, priv->numa_node), - priv->irq_info[i].mask); + priv->irq_info[vecidx].mask); if (IS_ENABLED(CONFIG_SMP) && - irq_set_affinity_hint(irq, priv->irq_info[i].mask)) + irq_set_affinity_hint(irq, priv->irq_info[vecidx].mask)) mlx5_core_warn(mdev, "irq_set_affinity_hint failed, irq 0x%.4x", irq); return 0; @@ -659,11 +660,12 @@ static int mlx5_irq_set_affinity_hint(struct mlx5_core_dev *mdev, int i) static void mlx5_irq_clear_affinity_hint(struct mlx5_core_dev *mdev, int i) { + int vecidx = MLX5_EQ_VEC_COMP_BASE + i; struct mlx5_priv *priv = &mdev->priv; - int irq = pci_irq_vector(mdev->pdev, MLX5_EQ_VEC_COMP_BASE + i); + int irq = pci_irq_vector(mdev->pdev, vecidx); irq_set_affinity_hint(irq, NULL); - free_cpumask_var(priv->irq_info[i].mask); + free_cpumask_var(priv->irq_info[vecidx].mask); } static int mlx5_irq_set_affinity_hints(struct mlx5_core_dev *mdev) diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index aa5963b5d38e..7d4ed995b4ce 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -1309,7 +1309,7 @@ enum { static inline const struct cpumask * mlx5_get_vector_affinity_hint(struct mlx5_core_dev *dev, int vector) { - return dev->priv.irq_info[vector].mask; + return dev->priv.irq_info[vector + MLX5_EQ_VEC_COMP_BASE].mask; } #endif /* MLX5_DRIVER_H */ -- cgit v1.2.3-55-g7522 From 4de45c758636c37efd313589f91c739f613fbe7d Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 19 Nov 2018 10:52:32 -0800 Subject: net/mlx5: EQ, Remove unused fields and structures Some fields and structures are not referenced nor used by the driver, remove them. Signed-off-by: Saeed Mahameed Reviewed-by: Leon Romanovsky Reviewed-by: Tariq Toukan Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 11 ----------- include/linux/mlx5/driver.h | 3 --- 2 files changed, 14 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index aeab0c4f60f4..fd5926daa0a6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -78,17 +78,6 @@ enum { (1ull << MLX5_EVENT_TYPE_SRQ_LAST_WQE) | \ (1ull << MLX5_EVENT_TYPE_SRQ_RQ_LIMIT)) -struct map_eq_in { - u64 mask; - u32 reserved; - u32 unmap_eqn; -}; - -struct cre_des_eq { - u8 reserved[15]; - u8 eqn; -}; - static int mlx5_cmd_destroy_eq(struct mlx5_core_dev *dev, u8 eqn) { u32 out[MLX5_ST_SZ_DW(destroy_eq_out)] = {0}; diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 7d4ed995b4ce..15cf6727a62d 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -398,7 +398,6 @@ struct mlx5_eq { unsigned int irqn; u8 eqn; int nent; - u64 mask; struct list_head list; int index; struct mlx5_rsc_debug *dbg; @@ -478,8 +477,6 @@ struct mlx5_core_srq { }; struct mlx5_eq_table { - void __iomem *update_ci; - void __iomem *update_arm_ci; struct list_head comp_eqs_list; struct mlx5_eq pages_eq; struct mlx5_eq async_eq; -- cgit v1.2.3-55-g7522 From 2883f352571b9b830561ca21b8a666936366a120 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 19 Nov 2018 10:52:33 -0800 Subject: net/mlx5: EQ, No need to store eq index as a field eq->index is used only for completion EQs and is assigned to be the completion eq index, it is used only when traversing the completion eqs list, and it can be calculated dynamically, thus remove the eq->index field. Signed-off-by: Saeed Mahameed Reviewed-by: Leon Romanovsky Reviewed-by: Tariq Toukan Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/main.c | 4 ++-- include/linux/mlx5/driver.h | 1 - 2 files changed, 2 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index d5cea0a36e6a..f5e6d375a8cc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -702,10 +702,11 @@ int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn, struct mlx5_eq_table *table = &dev->priv.eq_table; struct mlx5_eq *eq, *n; int err = -ENOENT; + int i = 0; spin_lock(&table->lock); list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) { - if (eq->index == vector) { + if (i++ == vector) { *eqn = eq->eqn; *irqn = eq->irqn; err = 0; @@ -797,7 +798,6 @@ static int alloc_comp_eqs(struct mlx5_core_dev *dev) goto clean; } mlx5_core_dbg(dev, "allocated completion EQN %d\n", eq->eqn); - eq->index = i; spin_lock(&table->lock); list_add_tail(&eq->list, &table->comp_eqs_list); spin_unlock(&table->lock); diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 15cf6727a62d..4b62d71825c1 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -399,7 +399,6 @@ struct mlx5_eq { u8 eqn; int nent; struct list_head list; - int index; struct mlx5_rsc_debug *dbg; enum mlx5_eq_type type; union { -- cgit v1.2.3-55-g7522 From aaa553a64438640ee4e41a2c1027c3435a75c0e7 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 19 Nov 2018 10:52:34 -0800 Subject: net/mlx5: EQ, Remove redundant completion EQ list lock Completion EQs list is only modified on driver load/unload, locking is not required, remove it. Signed-off-by: Saeed Mahameed Reviewed-by: Leon Romanovsky Reviewed-by: Tariq Toukan Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 2 -- drivers/net/ethernet/mellanox/mlx5/core/main.c | 17 +++-------------- include/linux/mlx5/driver.h | 3 --- 3 files changed, 3 insertions(+), 19 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index fd5926daa0a6..e75272503027 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -810,8 +810,6 @@ int mlx5_eq_init(struct mlx5_core_dev *dev) { int err; - spin_lock_init(&dev->priv.eq_table.lock); - err = mlx5_eq_debugfs_init(dev); return err; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index f5e6d375a8cc..f692c2a42130 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -704,7 +704,6 @@ int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn, int err = -ENOENT; int i = 0; - spin_lock(&table->lock); list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) { if (i++ == vector) { *eqn = eq->eqn; @@ -713,7 +712,6 @@ int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn, break; } } - spin_unlock(&table->lock); return err; } @@ -724,14 +722,11 @@ struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn) struct mlx5_eq_table *table = &dev->priv.eq_table; struct mlx5_eq *eq; - spin_lock(&table->lock); - list_for_each_entry(eq, &table->comp_eqs_list, list) - if (eq->eqn == eqn) { - spin_unlock(&table->lock); + list_for_each_entry(eq, &table->comp_eqs_list, list) { + if (eq->eqn == eqn) return eq; - } + } - spin_unlock(&table->lock); return ERR_PTR(-ENOENT); } @@ -747,17 +742,13 @@ static void free_comp_eqs(struct mlx5_core_dev *dev) dev->rmap = NULL; } #endif - spin_lock(&table->lock); list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) { list_del(&eq->list); - spin_unlock(&table->lock); if (mlx5_destroy_unmap_eq(dev, eq)) mlx5_core_warn(dev, "failed to destroy EQ 0x%x\n", eq->eqn); kfree(eq); - spin_lock(&table->lock); } - spin_unlock(&table->lock); } static int alloc_comp_eqs(struct mlx5_core_dev *dev) @@ -798,9 +789,7 @@ static int alloc_comp_eqs(struct mlx5_core_dev *dev) goto clean; } mlx5_core_dbg(dev, "allocated completion EQN %d\n", eq->eqn); - spin_lock(&table->lock); list_add_tail(&eq->list, &table->comp_eqs_list); - spin_unlock(&table->lock); } return 0; diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 4b62d71825c1..852e397c7624 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -484,9 +484,6 @@ struct mlx5_eq_table { struct mlx5_eq pfault_eq; #endif int num_comp_vectors; - /* protect EQs list - */ - spinlock_t lock; }; struct mlx5_uars_page { -- cgit v1.2.3-55-g7522 From ca828cb4686f1aece8d30541e11b8e21de1a7b0e Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 19 Nov 2018 10:52:35 -0800 Subject: net/mlx5: EQ, Move all EQ logic to eq.c Move completion EQs flows from main.c to eq.c, reasons: 1) It is where this logic belongs. 2) It will help centralize the EQ logic in one file for downstream refactoring, and future extensions/updates. Signed-off-by: Saeed Mahameed Reviewed-by: Leon Romanovsky Reviewed-by: Tariq Toukan Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 176 ++++++++++++++++++++ drivers/net/ethernet/mellanox/mlx5/core/main.c | 179 +-------------------- .../net/ethernet/mellanox/mlx5/core/mlx5_core.h | 2 + 3 files changed, 181 insertions(+), 176 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index e75272503027..4d79a4ccb758 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -820,6 +820,8 @@ void mlx5_eq_cleanup(struct mlx5_core_dev *dev) mlx5_eq_debugfs_cleanup(dev); } +/* Async EQs */ + int mlx5_start_eqs(struct mlx5_core_dev *dev) { struct mlx5_eq_table *table = &dev->priv.eq_table; @@ -953,12 +955,186 @@ int mlx5_core_eq_query(struct mlx5_core_dev *dev, struct mlx5_eq *eq, return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen); } +/* Completion EQs */ + +static int mlx5_irq_set_affinity_hint(struct mlx5_core_dev *mdev, int i) +{ + struct mlx5_priv *priv = &mdev->priv; + int vecidx = MLX5_EQ_VEC_COMP_BASE + i; + int irq = pci_irq_vector(mdev->pdev, vecidx); + + if (!zalloc_cpumask_var(&priv->irq_info[vecidx].mask, GFP_KERNEL)) { + mlx5_core_warn(mdev, "zalloc_cpumask_var failed"); + return -ENOMEM; + } + + cpumask_set_cpu(cpumask_local_spread(i, priv->numa_node), + priv->irq_info[vecidx].mask); + + if (IS_ENABLED(CONFIG_SMP) && + irq_set_affinity_hint(irq, priv->irq_info[vecidx].mask)) + mlx5_core_warn(mdev, "irq_set_affinity_hint failed, irq 0x%.4x", irq); + + return 0; +} + +static void mlx5_irq_clear_affinity_hint(struct mlx5_core_dev *mdev, int i) +{ + int vecidx = MLX5_EQ_VEC_COMP_BASE + i; + struct mlx5_priv *priv = &mdev->priv; + int irq = pci_irq_vector(mdev->pdev, vecidx); + + irq_set_affinity_hint(irq, NULL); + free_cpumask_var(priv->irq_info[vecidx].mask); +} + +static int mlx5_irq_set_affinity_hints(struct mlx5_core_dev *mdev) +{ + int err; + int i; + + for (i = 0; i < mdev->priv.eq_table.num_comp_vectors; i++) { + err = mlx5_irq_set_affinity_hint(mdev, i); + if (err) + goto err_out; + } + + return 0; + +err_out: + for (i--; i >= 0; i--) + mlx5_irq_clear_affinity_hint(mdev, i); + + return err; +} + +static void mlx5_irq_clear_affinity_hints(struct mlx5_core_dev *mdev) +{ + int i; + + for (i = 0; i < mdev->priv.eq_table.num_comp_vectors; i++) + mlx5_irq_clear_affinity_hint(mdev, i); +} + +void mlx5_free_comp_eqs(struct mlx5_core_dev *dev) +{ + struct mlx5_eq_table *table = &dev->priv.eq_table; + struct mlx5_eq *eq, *n; + + mlx5_irq_clear_affinity_hints(dev); + +#ifdef CONFIG_RFS_ACCEL + if (dev->rmap) { + free_irq_cpu_rmap(dev->rmap); + dev->rmap = NULL; + } +#endif + list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) { + list_del(&eq->list); + if (mlx5_destroy_unmap_eq(dev, eq)) + mlx5_core_warn(dev, "failed to destroy EQ 0x%x\n", + eq->eqn); + kfree(eq); + } +} + +int mlx5_alloc_comp_eqs(struct mlx5_core_dev *dev) +{ + struct mlx5_eq_table *table = &dev->priv.eq_table; + char name[MLX5_MAX_IRQ_NAME]; + struct mlx5_eq *eq; + int ncomp_vec; + int nent; + int err; + int i; + + INIT_LIST_HEAD(&table->comp_eqs_list); + ncomp_vec = table->num_comp_vectors; + nent = MLX5_COMP_EQ_SIZE; +#ifdef CONFIG_RFS_ACCEL + dev->rmap = alloc_irq_cpu_rmap(ncomp_vec); + if (!dev->rmap) + return -ENOMEM; +#endif + for (i = 0; i < ncomp_vec; i++) { + int vecidx = i + MLX5_EQ_VEC_COMP_BASE; + + eq = kzalloc(sizeof(*eq), GFP_KERNEL); + if (!eq) { + err = -ENOMEM; + goto clean; + } + +#ifdef CONFIG_RFS_ACCEL + irq_cpu_rmap_add(dev->rmap, pci_irq_vector(dev->pdev, vecidx)); +#endif + snprintf(name, MLX5_MAX_IRQ_NAME, "mlx5_comp%d", i); + err = mlx5_create_map_eq(dev, eq, vecidx, nent, 0, + name, MLX5_EQ_TYPE_COMP); + if (err) { + kfree(eq); + goto clean; + } + mlx5_core_dbg(dev, "allocated completion EQN %d\n", eq->eqn); + /* add tail, to keep the list ordered, for mlx5_vector2eqn to work */ + list_add_tail(&eq->list, &table->comp_eqs_list); + } + + err = mlx5_irq_set_affinity_hints(dev); + if (err) { + mlx5_core_err(dev, "Failed to alloc affinity hint cpumask\n"); + goto clean; + } + + return 0; + +clean: + mlx5_free_comp_eqs(dev); + return err; +} + +int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn, + unsigned int *irqn) +{ + struct mlx5_eq_table *table = &dev->priv.eq_table; + struct mlx5_eq *eq, *n; + int err = -ENOENT; + int i = 0; + + list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) { + if (i++ == vector) { + *eqn = eq->eqn; + *irqn = eq->irqn; + err = 0; + break; + } + } + + return err; +} +EXPORT_SYMBOL(mlx5_vector2eqn); + +struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn) +{ + struct mlx5_eq_table *table = &dev->priv.eq_table; + struct mlx5_eq *eq; + + list_for_each_entry(eq, &table->comp_eqs_list, list) { + if (eq->eqn == eqn) + return eq; + } + + return ERR_PTR(-ENOENT); +} + /* This function should only be called after mlx5_cmd_force_teardown_hca */ void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev) { struct mlx5_eq_table *table = &dev->priv.eq_table; struct mlx5_eq *eq; + mlx5_irq_clear_affinity_hints(dev); + #ifdef CONFIG_RFS_ACCEL if (dev->rmap) { free_irq_cpu_rmap(dev->rmap); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index f692c2a42130..244fec4b2ef2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -637,168 +637,6 @@ u64 mlx5_read_internal_timer(struct mlx5_core_dev *dev) return (u64)timer_l | (u64)timer_h1 << 32; } -static int mlx5_irq_set_affinity_hint(struct mlx5_core_dev *mdev, int i) -{ - struct mlx5_priv *priv = &mdev->priv; - int vecidx = MLX5_EQ_VEC_COMP_BASE + i; - int irq = pci_irq_vector(mdev->pdev, vecidx); - - if (!zalloc_cpumask_var(&priv->irq_info[vecidx].mask, GFP_KERNEL)) { - mlx5_core_warn(mdev, "zalloc_cpumask_var failed"); - return -ENOMEM; - } - - cpumask_set_cpu(cpumask_local_spread(i, priv->numa_node), - priv->irq_info[vecidx].mask); - - if (IS_ENABLED(CONFIG_SMP) && - irq_set_affinity_hint(irq, priv->irq_info[vecidx].mask)) - mlx5_core_warn(mdev, "irq_set_affinity_hint failed, irq 0x%.4x", irq); - - return 0; -} - -static void mlx5_irq_clear_affinity_hint(struct mlx5_core_dev *mdev, int i) -{ - int vecidx = MLX5_EQ_VEC_COMP_BASE + i; - struct mlx5_priv *priv = &mdev->priv; - int irq = pci_irq_vector(mdev->pdev, vecidx); - - irq_set_affinity_hint(irq, NULL); - free_cpumask_var(priv->irq_info[vecidx].mask); -} - -static int mlx5_irq_set_affinity_hints(struct mlx5_core_dev *mdev) -{ - int err; - int i; - - for (i = 0; i < mdev->priv.eq_table.num_comp_vectors; i++) { - err = mlx5_irq_set_affinity_hint(mdev, i); - if (err) - goto err_out; - } - - return 0; - -err_out: - for (i--; i >= 0; i--) - mlx5_irq_clear_affinity_hint(mdev, i); - - return err; -} - -static void mlx5_irq_clear_affinity_hints(struct mlx5_core_dev *mdev) -{ - int i; - - for (i = 0; i < mdev->priv.eq_table.num_comp_vectors; i++) - mlx5_irq_clear_affinity_hint(mdev, i); -} - -int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn, - unsigned int *irqn) -{ - struct mlx5_eq_table *table = &dev->priv.eq_table; - struct mlx5_eq *eq, *n; - int err = -ENOENT; - int i = 0; - - list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) { - if (i++ == vector) { - *eqn = eq->eqn; - *irqn = eq->irqn; - err = 0; - break; - } - } - - return err; -} -EXPORT_SYMBOL(mlx5_vector2eqn); - -struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn) -{ - struct mlx5_eq_table *table = &dev->priv.eq_table; - struct mlx5_eq *eq; - - list_for_each_entry(eq, &table->comp_eqs_list, list) { - if (eq->eqn == eqn) - return eq; - } - - - return ERR_PTR(-ENOENT); -} - -static void free_comp_eqs(struct mlx5_core_dev *dev) -{ - struct mlx5_eq_table *table = &dev->priv.eq_table; - struct mlx5_eq *eq, *n; - -#ifdef CONFIG_RFS_ACCEL - if (dev->rmap) { - free_irq_cpu_rmap(dev->rmap); - dev->rmap = NULL; - } -#endif - list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) { - list_del(&eq->list); - if (mlx5_destroy_unmap_eq(dev, eq)) - mlx5_core_warn(dev, "failed to destroy EQ 0x%x\n", - eq->eqn); - kfree(eq); - } -} - -static int alloc_comp_eqs(struct mlx5_core_dev *dev) -{ - struct mlx5_eq_table *table = &dev->priv.eq_table; - char name[MLX5_MAX_IRQ_NAME]; - struct mlx5_eq *eq; - int ncomp_vec; - int nent; - int err; - int i; - - INIT_LIST_HEAD(&table->comp_eqs_list); - ncomp_vec = table->num_comp_vectors; - nent = MLX5_COMP_EQ_SIZE; -#ifdef CONFIG_RFS_ACCEL - dev->rmap = alloc_irq_cpu_rmap(ncomp_vec); - if (!dev->rmap) - return -ENOMEM; -#endif - for (i = 0; i < ncomp_vec; i++) { - eq = kzalloc(sizeof(*eq), GFP_KERNEL); - if (!eq) { - err = -ENOMEM; - goto clean; - } - -#ifdef CONFIG_RFS_ACCEL - irq_cpu_rmap_add(dev->rmap, pci_irq_vector(dev->pdev, - MLX5_EQ_VEC_COMP_BASE + i)); -#endif - snprintf(name, MLX5_MAX_IRQ_NAME, "mlx5_comp%d", i); - err = mlx5_create_map_eq(dev, eq, - i + MLX5_EQ_VEC_COMP_BASE, nent, 0, - name, MLX5_EQ_TYPE_COMP); - if (err) { - kfree(eq); - goto clean; - } - mlx5_core_dbg(dev, "allocated completion EQN %d\n", eq->eqn); - list_add_tail(&eq->list, &table->comp_eqs_list); - } - - return 0; - -clean: - free_comp_eqs(dev); - return err; -} - static int mlx5_core_set_issi(struct mlx5_core_dev *dev) { u32 query_in[MLX5_ST_SZ_DW(query_issi_in)] = {0}; @@ -1177,18 +1015,12 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, goto err_fw_tracer; } - err = alloc_comp_eqs(dev); + err = mlx5_alloc_comp_eqs(dev); if (err) { dev_err(&pdev->dev, "Failed to alloc completion EQs\n"); goto err_comp_eqs; } - err = mlx5_irq_set_affinity_hints(dev); - if (err) { - dev_err(&pdev->dev, "Failed to alloc affinity hint cpumask\n"); - goto err_affinity_hints; - } - err = mlx5_fpga_device_start(dev); if (err) { dev_err(&pdev->dev, "fpga device start failed %d\n", err); @@ -1257,10 +1089,7 @@ err_ipsec_start: mlx5_fpga_device_stop(dev); err_fpga_start: - mlx5_irq_clear_affinity_hints(dev); - -err_affinity_hints: - free_comp_eqs(dev); + mlx5_free_comp_eqs(dev); err_comp_eqs: mlx5_fw_tracer_cleanup(dev->tracer); @@ -1331,8 +1160,7 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, mlx5_accel_ipsec_cleanup(dev); mlx5_accel_tls_cleanup(dev); mlx5_fpga_device_stop(dev); - mlx5_irq_clear_affinity_hints(dev); - free_comp_eqs(dev); + mlx5_free_comp_eqs(dev); mlx5_fw_tracer_cleanup(dev->tracer); mlx5_stop_eqs(dev); mlx5_put_uars_page(dev, priv->uar); @@ -1628,7 +1456,6 @@ succeed: * kexec. There is no need to cleanup the mlx5_core software * contexts. */ - mlx5_irq_clear_affinity_hints(dev); mlx5_core_eq_free_irqs(dev); return 0; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index 0594d0961cb3..22cff00faa5a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -136,6 +136,8 @@ int mlx5_core_eq_query(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u32 *out, int outlen); int mlx5_start_eqs(struct mlx5_core_dev *dev); void mlx5_stop_eqs(struct mlx5_core_dev *dev); +int mlx5_alloc_comp_eqs(struct mlx5_core_dev *dev); +void mlx5_free_comp_eqs(struct mlx5_core_dev *dev); /* This function should only be called after mlx5_cmd_force_teardown_hca */ void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev); struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn); -- cgit v1.2.3-55-g7522 From c8e21b3b576b78fe1b07522aea046af2634a24e8 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 19 Nov 2018 10:52:36 -0800 Subject: net/mlx5: EQ, Create all EQs in one place Instead of creating the EQ table in three steps at driver load, - allocate irq vectors - allocate async EQs - allocate completion EQs Gather all of the procedures into one function in eq.c and call it from driver load. This will help us reduce the EQ and EQ table private structures visibility to eq.c in downstream refactoring. Signed-off-by: Saeed Mahameed Reviewed-by: Leon Romanovsky Reviewed-by: Tariq Toukan Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/debugfs.c | 10 ++ drivers/net/ethernet/mellanox/mlx5/core/eq.c | 121 ++++++++++++++++----- drivers/net/ethernet/mellanox/mlx5/core/main.c | 81 ++------------ .../net/ethernet/mellanox/mlx5/core/mlx5_core.h | 8 +- 4 files changed, 116 insertions(+), 104 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c b/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c index 90fabd612b6c..b76766fb6c67 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c @@ -349,6 +349,16 @@ out: return param; } +static int mlx5_core_eq_query(struct mlx5_core_dev *dev, struct mlx5_eq *eq, + u32 *out, int outlen) +{ + u32 in[MLX5_ST_SZ_DW(query_eq_in)] = {}; + + MLX5_SET(query_eq_in, in, opcode, MLX5_CMD_OP_QUERY_EQ); + MLX5_SET(query_eq_in, in, eq_number, eq->eqn); + return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen); +} + static u64 eq_read_field(struct mlx5_core_dev *dev, struct mlx5_eq *eq, int index) { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 4d79a4ccb758..44ccd4206104 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -822,7 +822,7 @@ void mlx5_eq_cleanup(struct mlx5_core_dev *dev) /* Async EQs */ -int mlx5_start_eqs(struct mlx5_core_dev *dev) +static int create_async_eqs(struct mlx5_core_dev *dev) { struct mlx5_eq_table *table = &dev->priv.eq_table; u64 async_event_mask = MLX5_ASYNC_EVENT_MASK; @@ -914,7 +914,7 @@ err1: return err; } -void mlx5_stop_eqs(struct mlx5_core_dev *dev) +static void destroy_async_eqs(struct mlx5_core_dev *dev) { struct mlx5_eq_table *table = &dev->priv.eq_table; int err; @@ -945,19 +945,9 @@ void mlx5_stop_eqs(struct mlx5_core_dev *dev) err); } -int mlx5_core_eq_query(struct mlx5_core_dev *dev, struct mlx5_eq *eq, - u32 *out, int outlen) -{ - u32 in[MLX5_ST_SZ_DW(query_eq_in)] = {0}; - - MLX5_SET(query_eq_in, in, opcode, MLX5_CMD_OP_QUERY_EQ); - MLX5_SET(query_eq_in, in, eq_number, eq->eqn); - return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen); -} - /* Completion EQs */ -static int mlx5_irq_set_affinity_hint(struct mlx5_core_dev *mdev, int i) +static int set_comp_irq_affinity_hint(struct mlx5_core_dev *mdev, int i) { struct mlx5_priv *priv = &mdev->priv; int vecidx = MLX5_EQ_VEC_COMP_BASE + i; @@ -978,7 +968,7 @@ static int mlx5_irq_set_affinity_hint(struct mlx5_core_dev *mdev, int i) return 0; } -static void mlx5_irq_clear_affinity_hint(struct mlx5_core_dev *mdev, int i) +static void clear_comp_irq_affinity_hint(struct mlx5_core_dev *mdev, int i) { int vecidx = MLX5_EQ_VEC_COMP_BASE + i; struct mlx5_priv *priv = &mdev->priv; @@ -988,13 +978,13 @@ static void mlx5_irq_clear_affinity_hint(struct mlx5_core_dev *mdev, int i) free_cpumask_var(priv->irq_info[vecidx].mask); } -static int mlx5_irq_set_affinity_hints(struct mlx5_core_dev *mdev) +static int set_comp_irq_affinity_hints(struct mlx5_core_dev *mdev) { int err; int i; for (i = 0; i < mdev->priv.eq_table.num_comp_vectors; i++) { - err = mlx5_irq_set_affinity_hint(mdev, i); + err = set_comp_irq_affinity_hint(mdev, i); if (err) goto err_out; } @@ -1003,25 +993,25 @@ static int mlx5_irq_set_affinity_hints(struct mlx5_core_dev *mdev) err_out: for (i--; i >= 0; i--) - mlx5_irq_clear_affinity_hint(mdev, i); + clear_comp_irq_affinity_hint(mdev, i); return err; } -static void mlx5_irq_clear_affinity_hints(struct mlx5_core_dev *mdev) +static void clear_comp_irqs_affinity_hints(struct mlx5_core_dev *mdev) { int i; for (i = 0; i < mdev->priv.eq_table.num_comp_vectors; i++) - mlx5_irq_clear_affinity_hint(mdev, i); + clear_comp_irq_affinity_hint(mdev, i); } -void mlx5_free_comp_eqs(struct mlx5_core_dev *dev) +static void destroy_comp_eqs(struct mlx5_core_dev *dev) { struct mlx5_eq_table *table = &dev->priv.eq_table; struct mlx5_eq *eq, *n; - mlx5_irq_clear_affinity_hints(dev); + clear_comp_irqs_affinity_hints(dev); #ifdef CONFIG_RFS_ACCEL if (dev->rmap) { @@ -1038,7 +1028,7 @@ void mlx5_free_comp_eqs(struct mlx5_core_dev *dev) } } -int mlx5_alloc_comp_eqs(struct mlx5_core_dev *dev) +static int create_comp_eqs(struct mlx5_core_dev *dev) { struct mlx5_eq_table *table = &dev->priv.eq_table; char name[MLX5_MAX_IRQ_NAME]; @@ -1080,7 +1070,7 @@ int mlx5_alloc_comp_eqs(struct mlx5_core_dev *dev) list_add_tail(&eq->list, &table->comp_eqs_list); } - err = mlx5_irq_set_affinity_hints(dev); + err = set_comp_irq_affinity_hints(dev); if (err) { mlx5_core_err(dev, "Failed to alloc affinity hint cpumask\n"); goto clean; @@ -1089,7 +1079,7 @@ int mlx5_alloc_comp_eqs(struct mlx5_core_dev *dev) return 0; clean: - mlx5_free_comp_eqs(dev); + destroy_comp_eqs(dev); return err; } @@ -1133,7 +1123,7 @@ void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev) struct mlx5_eq_table *table = &dev->priv.eq_table; struct mlx5_eq *eq; - mlx5_irq_clear_affinity_hints(dev); + clear_comp_irqs_affinity_hints(dev); #ifdef CONFIG_RFS_ACCEL if (dev->rmap) { @@ -1153,3 +1143,84 @@ void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev) #endif pci_free_irq_vectors(dev->pdev); } + +static int alloc_irq_vectors(struct mlx5_core_dev *dev) +{ + struct mlx5_priv *priv = &dev->priv; + struct mlx5_eq_table *table = &priv->eq_table; + int num_eqs = MLX5_CAP_GEN(dev, max_num_eqs) ? + MLX5_CAP_GEN(dev, max_num_eqs) : + 1 << MLX5_CAP_GEN(dev, log_max_eq); + int nvec; + int err; + + nvec = MLX5_CAP_GEN(dev, num_ports) * num_online_cpus() + + MLX5_EQ_VEC_COMP_BASE; + nvec = min_t(int, nvec, num_eqs); + if (nvec <= MLX5_EQ_VEC_COMP_BASE) + return -ENOMEM; + + priv->irq_info = kcalloc(nvec, sizeof(*priv->irq_info), GFP_KERNEL); + if (!priv->irq_info) + return -ENOMEM; + + nvec = pci_alloc_irq_vectors(dev->pdev, MLX5_EQ_VEC_COMP_BASE + 1, + nvec, PCI_IRQ_MSIX); + if (nvec < 0) { + err = nvec; + goto err_free_irq_info; + } + + table->num_comp_vectors = nvec - MLX5_EQ_VEC_COMP_BASE; + + return 0; + +err_free_irq_info: + kfree(priv->irq_info); + return err; +} + +static void free_irq_vectors(struct mlx5_core_dev *dev) +{ + struct mlx5_priv *priv = &dev->priv; + + pci_free_irq_vectors(dev->pdev); + kfree(priv->irq_info); +} + +int mlx5_eq_table_create(struct mlx5_core_dev *dev) +{ + int err; + + err = alloc_irq_vectors(dev); + if (err) { + mlx5_core_err(dev, "alloc irq vectors failed\n"); + return err; + } + + err = create_async_eqs(dev); + if (err) { + mlx5_core_err(dev, "Failed to create async EQs\n"); + goto err_async_eqs; + } + + err = create_comp_eqs(dev); + if (err) { + mlx5_core_err(dev, "Failed to create completion EQs\n"); + goto err_comp_eqs; + } + + return 0; +err_comp_eqs: + destroy_async_eqs(dev); +err_async_eqs: + free_irq_vectors(dev); + return err; +} + +void mlx5_eq_table_destroy(struct mlx5_core_dev *dev) +{ + destroy_comp_eqs(dev); + destroy_async_eqs(dev); + free_irq_vectors(dev); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 244fec4b2ef2..21cc9bbc2563 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -319,51 +319,6 @@ static void release_bar(struct pci_dev *pdev) pci_release_regions(pdev); } -static int mlx5_alloc_irq_vectors(struct mlx5_core_dev *dev) -{ - struct mlx5_priv *priv = &dev->priv; - struct mlx5_eq_table *table = &priv->eq_table; - int num_eqs = MLX5_CAP_GEN(dev, max_num_eqs) ? - MLX5_CAP_GEN(dev, max_num_eqs) : - 1 << MLX5_CAP_GEN(dev, log_max_eq); - int nvec; - int err; - - nvec = MLX5_CAP_GEN(dev, num_ports) * num_online_cpus() + - MLX5_EQ_VEC_COMP_BASE; - nvec = min_t(int, nvec, num_eqs); - if (nvec <= MLX5_EQ_VEC_COMP_BASE) - return -ENOMEM; - - priv->irq_info = kcalloc(nvec, sizeof(*priv->irq_info), GFP_KERNEL); - if (!priv->irq_info) - return -ENOMEM; - - nvec = pci_alloc_irq_vectors(dev->pdev, - MLX5_EQ_VEC_COMP_BASE + 1, nvec, - PCI_IRQ_MSIX); - if (nvec < 0) { - err = nvec; - goto err_free_irq_info; - } - - table->num_comp_vectors = nvec - MLX5_EQ_VEC_COMP_BASE; - - return 0; - -err_free_irq_info: - kfree(priv->irq_info); - return err; -} - -static void mlx5_free_irq_vectors(struct mlx5_core_dev *dev) -{ - struct mlx5_priv *priv = &dev->priv; - - pci_free_irq_vectors(dev->pdev); - kfree(priv->irq_info); -} - struct mlx5_reg_host_endianness { u8 he; u8 rsvd[15]; @@ -990,23 +945,17 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, } } - err = mlx5_alloc_irq_vectors(dev); - if (err) { - dev_err(&pdev->dev, "alloc irq vectors failed\n"); - goto err_cleanup_once; - } - dev->priv.uar = mlx5_get_uars_page(dev); if (IS_ERR(dev->priv.uar)) { dev_err(&pdev->dev, "Failed allocating uar, aborting\n"); err = PTR_ERR(dev->priv.uar); - goto err_disable_msix; + goto err_get_uars; } - err = mlx5_start_eqs(dev); + err = mlx5_eq_table_create(dev); if (err) { - dev_err(&pdev->dev, "Failed to start pages and async EQs\n"); - goto err_put_uars; + dev_err(&pdev->dev, "Failed to create EQs\n"); + goto err_eq_table; } err = mlx5_fw_tracer_init(dev->tracer); @@ -1015,12 +964,6 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, goto err_fw_tracer; } - err = mlx5_alloc_comp_eqs(dev); - if (err) { - dev_err(&pdev->dev, "Failed to alloc completion EQs\n"); - goto err_comp_eqs; - } - err = mlx5_fpga_device_start(dev); if (err) { dev_err(&pdev->dev, "fpga device start failed %d\n", err); @@ -1089,21 +1032,15 @@ err_ipsec_start: mlx5_fpga_device_stop(dev); err_fpga_start: - mlx5_free_comp_eqs(dev); - -err_comp_eqs: mlx5_fw_tracer_cleanup(dev->tracer); err_fw_tracer: - mlx5_stop_eqs(dev); + mlx5_eq_table_destroy(dev); -err_put_uars: +err_eq_table: mlx5_put_uars_page(dev, priv->uar); -err_disable_msix: - mlx5_free_irq_vectors(dev); - -err_cleanup_once: +err_get_uars: if (boot) mlx5_cleanup_once(dev); @@ -1160,11 +1097,9 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, mlx5_accel_ipsec_cleanup(dev); mlx5_accel_tls_cleanup(dev); mlx5_fpga_device_stop(dev); - mlx5_free_comp_eqs(dev); mlx5_fw_tracer_cleanup(dev->tracer); - mlx5_stop_eqs(dev); + mlx5_eq_table_destroy(dev); mlx5_put_uars_page(dev, priv->uar); - mlx5_free_irq_vectors(dev); if (cleanup) mlx5_cleanup_once(dev); mlx5_stop_health_poll(dev, cleanup); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index 22cff00faa5a..3fa6d26875fe 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -132,12 +132,8 @@ int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, int mlx5_destroy_unmap_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq); int mlx5_eq_add_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq); int mlx5_eq_del_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq); -int mlx5_core_eq_query(struct mlx5_core_dev *dev, struct mlx5_eq *eq, - u32 *out, int outlen); -int mlx5_start_eqs(struct mlx5_core_dev *dev); -void mlx5_stop_eqs(struct mlx5_core_dev *dev); -int mlx5_alloc_comp_eqs(struct mlx5_core_dev *dev); -void mlx5_free_comp_eqs(struct mlx5_core_dev *dev); +int mlx5_eq_table_create(struct mlx5_core_dev *dev); +void mlx5_eq_table_destroy(struct mlx5_core_dev *dev); /* This function should only be called after mlx5_cmd_force_teardown_hca */ void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev); struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn); -- cgit v1.2.3-55-g7522 From d674a9aa434409826b2408609be493739e61e6f6 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 19 Nov 2018 10:52:37 -0800 Subject: net/mlx5: EQ, irq_info and rmap belong to eq_table irq_info and rmap are EQ properties of the driver, and only needed for EQ objects, move them to the eq_table EQs database structure. Signed-off-by: Saeed Mahameed Reviewed-by: Leon Romanovsky Reviewed-by: Tariq Toukan Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 4 +-- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 40 ++++++++++++----------- include/linux/mlx5/driver.h | 10 +++--- 3 files changed, 28 insertions(+), 26 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 2839c30dd3a0..32ea47c28324 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -1760,7 +1760,7 @@ static void mlx5e_close_cq(struct mlx5e_cq *cq) static int mlx5e_get_cpu(struct mlx5e_priv *priv, int ix) { - return cpumask_first(priv->mdev->priv.irq_info[ix + MLX5_EQ_VEC_COMP_BASE].mask); + return cpumask_first(priv->mdev->priv.eq_table.irq_info[ix + MLX5_EQ_VEC_COMP_BASE].mask); } static int mlx5e_open_tx_cqs(struct mlx5e_channel *c, @@ -4960,7 +4960,7 @@ int mlx5e_netdev_init(struct net_device *netdev, netif_carrier_off(netdev); #ifdef CONFIG_MLX5_EN_ARFS - netdev->rx_cpu_rmap = mdev->rmap; + netdev->rx_cpu_rmap = mdev->priv.eq_table.rmap; #endif return 0; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 44ccd4206104..70f62f10065e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -694,7 +694,7 @@ int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, if (err) goto err_in; - snprintf(priv->irq_info[vecidx].name, MLX5_MAX_IRQ_NAME, "%s@pci:%s", + snprintf(priv->eq_table.irq_info[vecidx].name, MLX5_MAX_IRQ_NAME, "%s@pci:%s", name, pci_name(dev->pdev)); eq->eqn = MLX5_GET(create_eq_out, out, eq_number); @@ -702,7 +702,7 @@ int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, eq->dev = dev; eq->doorbell = priv->uar->map + MLX5_EQ_DOORBEL_OFFSET; err = request_irq(eq->irqn, handler, 0, - priv->irq_info[vecidx].name, eq); + priv->eq_table.irq_info[vecidx].name, eq); if (err) goto err_eq; @@ -952,17 +952,18 @@ static int set_comp_irq_affinity_hint(struct mlx5_core_dev *mdev, int i) struct mlx5_priv *priv = &mdev->priv; int vecidx = MLX5_EQ_VEC_COMP_BASE + i; int irq = pci_irq_vector(mdev->pdev, vecidx); + struct mlx5_irq_info *irq_info = &priv->eq_table.irq_info[vecidx]; - if (!zalloc_cpumask_var(&priv->irq_info[vecidx].mask, GFP_KERNEL)) { + if (!zalloc_cpumask_var(&irq_info->mask, GFP_KERNEL)) { mlx5_core_warn(mdev, "zalloc_cpumask_var failed"); return -ENOMEM; } cpumask_set_cpu(cpumask_local_spread(i, priv->numa_node), - priv->irq_info[vecidx].mask); + irq_info->mask); if (IS_ENABLED(CONFIG_SMP) && - irq_set_affinity_hint(irq, priv->irq_info[vecidx].mask)) + irq_set_affinity_hint(irq, irq_info->mask)) mlx5_core_warn(mdev, "irq_set_affinity_hint failed, irq 0x%.4x", irq); return 0; @@ -973,9 +974,10 @@ static void clear_comp_irq_affinity_hint(struct mlx5_core_dev *mdev, int i) int vecidx = MLX5_EQ_VEC_COMP_BASE + i; struct mlx5_priv *priv = &mdev->priv; int irq = pci_irq_vector(mdev->pdev, vecidx); + struct mlx5_irq_info *irq_info = &priv->eq_table.irq_info[vecidx]; irq_set_affinity_hint(irq, NULL); - free_cpumask_var(priv->irq_info[vecidx].mask); + free_cpumask_var(irq_info->mask); } static int set_comp_irq_affinity_hints(struct mlx5_core_dev *mdev) @@ -1014,9 +1016,9 @@ static void destroy_comp_eqs(struct mlx5_core_dev *dev) clear_comp_irqs_affinity_hints(dev); #ifdef CONFIG_RFS_ACCEL - if (dev->rmap) { - free_irq_cpu_rmap(dev->rmap); - dev->rmap = NULL; + if (table->rmap) { + free_irq_cpu_rmap(table->rmap); + table->rmap = NULL; } #endif list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) { @@ -1042,8 +1044,8 @@ static int create_comp_eqs(struct mlx5_core_dev *dev) ncomp_vec = table->num_comp_vectors; nent = MLX5_COMP_EQ_SIZE; #ifdef CONFIG_RFS_ACCEL - dev->rmap = alloc_irq_cpu_rmap(ncomp_vec); - if (!dev->rmap) + table->rmap = alloc_irq_cpu_rmap(ncomp_vec); + if (!table->rmap) return -ENOMEM; #endif for (i = 0; i < ncomp_vec; i++) { @@ -1056,7 +1058,7 @@ static int create_comp_eqs(struct mlx5_core_dev *dev) } #ifdef CONFIG_RFS_ACCEL - irq_cpu_rmap_add(dev->rmap, pci_irq_vector(dev->pdev, vecidx)); + irq_cpu_rmap_add(table->rmap, pci_irq_vector(dev->pdev, vecidx)); #endif snprintf(name, MLX5_MAX_IRQ_NAME, "mlx5_comp%d", i); err = mlx5_create_map_eq(dev, eq, vecidx, nent, 0, @@ -1126,9 +1128,9 @@ void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev) clear_comp_irqs_affinity_hints(dev); #ifdef CONFIG_RFS_ACCEL - if (dev->rmap) { - free_irq_cpu_rmap(dev->rmap); - dev->rmap = NULL; + if (table->rmap) { + free_irq_cpu_rmap(table->rmap); + table->rmap = NULL; } #endif list_for_each_entry(eq, &table->comp_eqs_list, list) @@ -1160,8 +1162,8 @@ static int alloc_irq_vectors(struct mlx5_core_dev *dev) if (nvec <= MLX5_EQ_VEC_COMP_BASE) return -ENOMEM; - priv->irq_info = kcalloc(nvec, sizeof(*priv->irq_info), GFP_KERNEL); - if (!priv->irq_info) + table->irq_info = kcalloc(nvec, sizeof(*table->irq_info), GFP_KERNEL); + if (!table->irq_info) return -ENOMEM; nvec = pci_alloc_irq_vectors(dev->pdev, MLX5_EQ_VEC_COMP_BASE + 1, @@ -1176,7 +1178,7 @@ static int alloc_irq_vectors(struct mlx5_core_dev *dev) return 0; err_free_irq_info: - kfree(priv->irq_info); + kfree(table->irq_info); return err; } @@ -1185,7 +1187,7 @@ static void free_irq_vectors(struct mlx5_core_dev *dev) struct mlx5_priv *priv = &dev->priv; pci_free_irq_vectors(dev->pdev); - kfree(priv->irq_info); + kfree(priv->eq_table.irq_info); } int mlx5_eq_table_create(struct mlx5_core_dev *dev) diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 852e397c7624..dcc3f7aa8572 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -484,6 +484,10 @@ struct mlx5_eq_table { struct mlx5_eq pfault_eq; #endif int num_comp_vectors; + struct mlx5_irq_info *irq_info; +#ifdef CONFIG_RFS_ACCEL + struct cpu_rmap *rmap; +#endif }; struct mlx5_uars_page { @@ -640,7 +644,6 @@ struct mlx5_port_module_event_stats { struct mlx5_priv { char name[MLX5_MAX_NAME_LEN]; struct mlx5_eq_table eq_table; - struct mlx5_irq_info *irq_info; /* pages stuff */ struct workqueue_struct *pg_wq; @@ -851,9 +854,6 @@ struct mlx5_core_dev { } roce; #ifdef CONFIG_MLX5_FPGA struct mlx5_fpga_device *fpga; -#endif -#ifdef CONFIG_RFS_ACCEL - struct cpu_rmap *rmap; #endif struct mlx5_clock clock; struct mlx5_ib_clock_info *clock_info; @@ -1302,7 +1302,7 @@ enum { static inline const struct cpumask * mlx5_get_vector_affinity_hint(struct mlx5_core_dev *dev, int vector) { - return dev->priv.irq_info[vector + MLX5_EQ_VEC_COMP_BASE].mask; + return dev->priv.eq_table.irq_info[vector + MLX5_EQ_VEC_COMP_BASE].mask; } #endif /* MLX5_DRIVER_H */ -- cgit v1.2.3-55-g7522 From f2f3df5501391bc784c8462dc97d989c2194fb74 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 19 Nov 2018 10:52:38 -0800 Subject: net/mlx5: EQ, Privatize eq_table and friends Move unnecessary EQ table structures and declaration from the public include/linux/mlx5/driver.h into the private area of mlx5_core and into eq.c/eq.h. Introduce new mlx5 EQ APIs: mlx5_comp_vectors_count(dev); mlx5_comp_irq_get_affinity_mask(dev, vector); And use them from mlx5_ib or mlx5e netdevice instead of direct access to mlx5_core internal structures. Signed-off-by: Saeed Mahameed Reviewed-by: Leon Romanovsky Reviewed-by: Tariq Toukan Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/mlx5/main.c | 5 +- drivers/net/ethernet/mellanox/mlx5/core/cq.c | 5 +- drivers/net/ethernet/mellanox/mlx5/core/debugfs.c | 1 + drivers/net/ethernet/mellanox/mlx5/core/en.h | 3 +- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 10 +- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 102 ++++++++++++++++----- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 1 + drivers/net/ethernet/mellanox/mlx5/core/health.c | 1 + drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h | 77 ++++++++++++++++ drivers/net/ethernet/mellanox/mlx5/core/main.c | 7 +- .../net/ethernet/mellanox/mlx5/core/mlx5_core.h | 15 --- include/linux/mlx5/driver.h | 87 +----------------- 12 files changed, 179 insertions(+), 135 deletions(-) create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index e9c428071df3..6fbc0cba1bac 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -5337,7 +5337,7 @@ mlx5_ib_get_vector_affinity(struct ib_device *ibdev, int comp_vector) { struct mlx5_ib_dev *dev = to_mdev(ibdev); - return mlx5_get_vector_affinity_hint(dev->mdev, comp_vector); + return mlx5_comp_irq_get_affinity_mask(dev->mdev, comp_vector); } /* The mlx5_ib_multiport_mutex should be held when calling this function */ @@ -5701,8 +5701,7 @@ int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev) dev->ib_dev.node_type = RDMA_NODE_IB_CA; dev->ib_dev.local_dma_lkey = 0 /* not supported for now */; dev->ib_dev.phys_port_cnt = dev->num_ports; - dev->ib_dev.num_comp_vectors = - dev->mdev->priv.eq_table.num_comp_vectors; + dev->ib_dev.num_comp_vectors = mlx5_comp_vectors_count(mdev); dev->ib_dev.dev.parent = &mdev->pdev->dev; mutex_init(&dev->cap_mask_mutex); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cq.c b/drivers/net/ethernet/mellanox/mlx5/core/cq.c index 4b85abb5c9f7..6e55d2f37c6d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/cq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/cq.c @@ -38,6 +38,7 @@ #include #include #include "mlx5_core.h" +#include "lib/eq.h" #define TASKLET_MAX_TIME 2 #define TASKLET_MAX_TIME_JIFFIES msecs_to_jiffies(TASKLET_MAX_TIME) @@ -124,7 +125,7 @@ int mlx5_core_create_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq, goto err_cmd; /* Add to async EQ CQ tree to recv async events */ - err = mlx5_eq_add_cq(&dev->priv.eq_table.async_eq, cq); + err = mlx5_eq_add_cq(mlx5_get_async_eq(dev), cq); if (err) goto err_cq_add; @@ -157,7 +158,7 @@ int mlx5_core_destroy_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq) u32 in[MLX5_ST_SZ_DW(destroy_cq_in)] = {0}; int err; - err = mlx5_eq_del_cq(&dev->priv.eq_table.async_eq, cq); + err = mlx5_eq_del_cq(mlx5_get_async_eq(dev), cq); if (err) return err; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c b/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c index b76766fb6c67..a11e22d0b0cc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c @@ -36,6 +36,7 @@ #include #include #include "mlx5_core.h" +#include "lib/eq.h" enum { QP_PID, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index d7fbd5b6ac95..aea74856c702 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -178,8 +178,7 @@ static inline int mlx5e_get_max_num_channels(struct mlx5_core_dev *mdev) { return is_kdump_kernel() ? MLX5E_MIN_NUM_CHANNELS : - min_t(int, mdev->priv.eq_table.num_comp_vectors, - MLX5E_MAX_NUM_CHANNELS); + min_t(int, mlx5_comp_vectors_count(mdev), MLX5E_MAX_NUM_CHANNELS); } /* Use this function to get max num channels after netdev was created */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 32ea47c28324..c23caade31bf 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -49,6 +49,7 @@ #include "lib/clock.h" #include "en/port.h" #include "en/xdp.h" +#include "lib/eq.h" struct mlx5e_rq_param { u32 rqc[MLX5_ST_SZ_DW(rqc)]; @@ -1758,11 +1759,6 @@ static void mlx5e_close_cq(struct mlx5e_cq *cq) mlx5e_free_cq(cq); } -static int mlx5e_get_cpu(struct mlx5e_priv *priv, int ix) -{ - return cpumask_first(priv->mdev->priv.eq_table.irq_info[ix + MLX5_EQ_VEC_COMP_BASE].mask); -} - static int mlx5e_open_tx_cqs(struct mlx5e_channel *c, struct mlx5e_params *params, struct mlx5e_channel_param *cparam) @@ -1913,9 +1909,9 @@ static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix, struct mlx5e_channel_param *cparam, struct mlx5e_channel **cp) { + int cpu = cpumask_first(mlx5_comp_irq_get_affinity_mask(priv->mdev, ix)); struct net_dim_cq_moder icocq_moder = {0, 0}; struct net_device *netdev = priv->netdev; - int cpu = mlx5e_get_cpu(priv, ix); struct mlx5e_channel *c; unsigned int irq; int err; @@ -4960,7 +4956,7 @@ int mlx5e_netdev_init(struct net_device *netdev, netif_carrier_off(netdev); #ifdef CONFIG_MLX5_EN_ARFS - netdev->rx_cpu_rmap = mdev->priv.eq_table.rmap; + netdev->rx_cpu_rmap = mlx5_eq_table_get_rmap(mdev); #endif return 0; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 70f62f10065e..32ce20221c44 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -38,6 +38,7 @@ #include #endif #include "mlx5_core.h" +#include "lib/eq.h" #include "fpga/core.h" #include "eswitch.h" #include "lib/clock.h" @@ -65,6 +66,26 @@ enum { MLX5_EQ_DOORBEL_OFFSET = 0x40, }; +struct mlx5_irq_info { + cpumask_var_t mask; + char name[MLX5_MAX_IRQ_NAME]; +}; + +struct mlx5_eq_table { + struct list_head comp_eqs_list; + struct mlx5_eq pages_eq; + struct mlx5_eq async_eq; + struct mlx5_eq cmd_eq; +#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING + struct mlx5_eq pfault_eq; +#endif + int num_comp_vectors; + struct mlx5_irq_info *irq_info; +#ifdef CONFIG_RFS_ACCEL + struct cpu_rmap *rmap; +#endif +}; + #define MLX5_ASYNC_EVENT_MASK ((1ull << MLX5_EVENT_TYPE_PATH_MIG) | \ (1ull << MLX5_EVENT_TYPE_COMM_EST) | \ (1ull << MLX5_EVENT_TYPE_SQ_DRAINED) | \ @@ -633,10 +654,11 @@ static void init_eq_buf(struct mlx5_eq *eq) } } -int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, - int nent, u64 mask, const char *name, - enum mlx5_eq_type type) +static int +mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, + int nent, u64 mask, const char *name, enum mlx5_eq_type type) { + struct mlx5_eq_table *eq_table = dev->priv.eq_table; struct mlx5_cq_table *cq_table = &eq->cq_table; u32 out[MLX5_ST_SZ_DW(create_eq_out)] = {0}; struct mlx5_priv *priv = &dev->priv; @@ -694,7 +716,7 @@ int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, if (err) goto err_in; - snprintf(priv->eq_table.irq_info[vecidx].name, MLX5_MAX_IRQ_NAME, "%s@pci:%s", + snprintf(eq_table->irq_info[vecidx].name, MLX5_MAX_IRQ_NAME, "%s@pci:%s", name, pci_name(dev->pdev)); eq->eqn = MLX5_GET(create_eq_out, out, eq_number); @@ -702,7 +724,7 @@ int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, eq->dev = dev; eq->doorbell = priv->uar->map + MLX5_EQ_DOORBEL_OFFSET; err = request_irq(eq->irqn, handler, 0, - priv->eq_table.irq_info[vecidx].name, eq); + eq_table->irq_info[vecidx].name, eq); if (err) goto err_eq; @@ -746,7 +768,7 @@ err_buf: return err; } -int mlx5_destroy_unmap_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq) +static int mlx5_destroy_unmap_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq) { int err; @@ -806,25 +828,35 @@ int mlx5_eq_del_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq) return 0; } -int mlx5_eq_init(struct mlx5_core_dev *dev) +int mlx5_eq_table_init(struct mlx5_core_dev *dev) { + struct mlx5_eq_table *eq_table; int err; + eq_table = kvzalloc(sizeof(*eq_table), GFP_KERNEL); + if (!eq_table) + return -ENOMEM; + + dev->priv.eq_table = eq_table; + err = mlx5_eq_debugfs_init(dev); + if (err) + kvfree(eq_table); return err; } -void mlx5_eq_cleanup(struct mlx5_core_dev *dev) +void mlx5_eq_table_cleanup(struct mlx5_core_dev *dev) { mlx5_eq_debugfs_cleanup(dev); + kvfree(dev->priv.eq_table); } /* Async EQs */ static int create_async_eqs(struct mlx5_core_dev *dev) { - struct mlx5_eq_table *table = &dev->priv.eq_table; + struct mlx5_eq_table *table = dev->priv.eq_table; u64 async_event_mask = MLX5_ASYNC_EVENT_MASK; int err; @@ -916,7 +948,7 @@ err1: static void destroy_async_eqs(struct mlx5_core_dev *dev) { - struct mlx5_eq_table *table = &dev->priv.eq_table; + struct mlx5_eq_table *table = dev->priv.eq_table; int err; #ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING @@ -945,6 +977,11 @@ static void destroy_async_eqs(struct mlx5_core_dev *dev) err); } +struct mlx5_eq *mlx5_get_async_eq(struct mlx5_core_dev *dev) +{ + return &dev->priv.eq_table->async_eq; +} + /* Completion EQs */ static int set_comp_irq_affinity_hint(struct mlx5_core_dev *mdev, int i) @@ -952,7 +989,7 @@ static int set_comp_irq_affinity_hint(struct mlx5_core_dev *mdev, int i) struct mlx5_priv *priv = &mdev->priv; int vecidx = MLX5_EQ_VEC_COMP_BASE + i; int irq = pci_irq_vector(mdev->pdev, vecidx); - struct mlx5_irq_info *irq_info = &priv->eq_table.irq_info[vecidx]; + struct mlx5_irq_info *irq_info = &priv->eq_table->irq_info[vecidx]; if (!zalloc_cpumask_var(&irq_info->mask, GFP_KERNEL)) { mlx5_core_warn(mdev, "zalloc_cpumask_var failed"); @@ -974,7 +1011,7 @@ static void clear_comp_irq_affinity_hint(struct mlx5_core_dev *mdev, int i) int vecidx = MLX5_EQ_VEC_COMP_BASE + i; struct mlx5_priv *priv = &mdev->priv; int irq = pci_irq_vector(mdev->pdev, vecidx); - struct mlx5_irq_info *irq_info = &priv->eq_table.irq_info[vecidx]; + struct mlx5_irq_info *irq_info = &priv->eq_table->irq_info[vecidx]; irq_set_affinity_hint(irq, NULL); free_cpumask_var(irq_info->mask); @@ -985,7 +1022,7 @@ static int set_comp_irq_affinity_hints(struct mlx5_core_dev *mdev) int err; int i; - for (i = 0; i < mdev->priv.eq_table.num_comp_vectors; i++) { + for (i = 0; i < mdev->priv.eq_table->num_comp_vectors; i++) { err = set_comp_irq_affinity_hint(mdev, i); if (err) goto err_out; @@ -1004,13 +1041,13 @@ static void clear_comp_irqs_affinity_hints(struct mlx5_core_dev *mdev) { int i; - for (i = 0; i < mdev->priv.eq_table.num_comp_vectors; i++) + for (i = 0; i < mdev->priv.eq_table->num_comp_vectors; i++) clear_comp_irq_affinity_hint(mdev, i); } static void destroy_comp_eqs(struct mlx5_core_dev *dev) { - struct mlx5_eq_table *table = &dev->priv.eq_table; + struct mlx5_eq_table *table = dev->priv.eq_table; struct mlx5_eq *eq, *n; clear_comp_irqs_affinity_hints(dev); @@ -1032,7 +1069,7 @@ static void destroy_comp_eqs(struct mlx5_core_dev *dev) static int create_comp_eqs(struct mlx5_core_dev *dev) { - struct mlx5_eq_table *table = &dev->priv.eq_table; + struct mlx5_eq_table *table = dev->priv.eq_table; char name[MLX5_MAX_IRQ_NAME]; struct mlx5_eq *eq; int ncomp_vec; @@ -1088,7 +1125,7 @@ clean: int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn, unsigned int *irqn) { - struct mlx5_eq_table *table = &dev->priv.eq_table; + struct mlx5_eq_table *table = dev->priv.eq_table; struct mlx5_eq *eq, *n; int err = -ENOENT; int i = 0; @@ -1106,9 +1143,32 @@ int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn, } EXPORT_SYMBOL(mlx5_vector2eqn); +unsigned int mlx5_comp_vectors_count(struct mlx5_core_dev *dev) +{ + return dev->priv.eq_table->num_comp_vectors; +} +EXPORT_SYMBOL(mlx5_comp_vectors_count); + +struct cpumask * +mlx5_comp_irq_get_affinity_mask(struct mlx5_core_dev *dev, int vector) +{ + /* TODO: consider irq_get_affinity_mask(irq) */ + return dev->priv.eq_table->irq_info[vector + MLX5_EQ_VEC_COMP_BASE].mask; +} +EXPORT_SYMBOL(mlx5_comp_irq_get_affinity_mask); + +struct cpu_rmap *mlx5_eq_table_get_rmap(struct mlx5_core_dev *dev) +{ +#ifdef CONFIG_RFS_ACCEL + return dev->priv.eq_table->rmap; +#else + return NULL; +#endif +} + struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn) { - struct mlx5_eq_table *table = &dev->priv.eq_table; + struct mlx5_eq_table *table = dev->priv.eq_table; struct mlx5_eq *eq; list_for_each_entry(eq, &table->comp_eqs_list, list) { @@ -1122,7 +1182,7 @@ struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn) /* This function should only be called after mlx5_cmd_force_teardown_hca */ void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev) { - struct mlx5_eq_table *table = &dev->priv.eq_table; + struct mlx5_eq_table *table = dev->priv.eq_table; struct mlx5_eq *eq; clear_comp_irqs_affinity_hints(dev); @@ -1149,7 +1209,7 @@ void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev) static int alloc_irq_vectors(struct mlx5_core_dev *dev) { struct mlx5_priv *priv = &dev->priv; - struct mlx5_eq_table *table = &priv->eq_table; + struct mlx5_eq_table *table = priv->eq_table; int num_eqs = MLX5_CAP_GEN(dev, max_num_eqs) ? MLX5_CAP_GEN(dev, max_num_eqs) : 1 << MLX5_CAP_GEN(dev, log_max_eq); @@ -1187,7 +1247,7 @@ static void free_irq_vectors(struct mlx5_core_dev *dev) struct mlx5_priv *priv = &dev->priv; pci_free_irq_vectors(dev->pdev); - kfree(priv->eq_table.irq_info); + kfree(priv->eq_table->irq_info); } int mlx5_eq_table_create(struct mlx5_core_dev *dev) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index d004957328f9..324606227b1a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -38,6 +38,7 @@ #include "mlx5_core.h" #include "eswitch.h" #include "fs_core.h" +#include "lib/eq.h" #define UPLINK_VPORT 0xFFFF diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c index 43118de8ee99..b5be6f0b9ed5 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/health.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c @@ -38,6 +38,7 @@ #include #include #include "mlx5_core.h" +#include "lib/eq.h" enum { MLX5_HEALTH_POLL_INTERVAL = 2 * HZ, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h new file mode 100644 index 000000000000..48ee37797b3f --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h @@ -0,0 +1,77 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* Copyright (c) 2018 Mellanox Technologies */ + +#ifndef __LIB_MLX5_EQ_H__ +#define __LIB_MLX5_EQ_H__ +#include + +#define MLX5_MAX_IRQ_NAME (32) + +enum { + MLX5_EQ_VEC_PAGES = 0, + MLX5_EQ_VEC_CMD = 1, + MLX5_EQ_VEC_ASYNC = 2, + MLX5_EQ_VEC_PFAULT = 3, + MLX5_EQ_VEC_COMP_BASE, +}; + +struct mlx5_eq_tasklet { + struct list_head list; + struct list_head process_list; + struct tasklet_struct task; + spinlock_t lock; /* lock completion tasklet list */ +}; + +struct mlx5_eq_pagefault { + struct work_struct work; + spinlock_t lock; /* Pagefaults spinlock */ + struct workqueue_struct *wq; + mempool_t *pool; +}; + +struct mlx5_cq_table { + spinlock_t lock; /* protect radix tree */ + struct radix_tree_root tree; +}; + +struct mlx5_eq { + struct mlx5_core_dev *dev; + struct mlx5_cq_table cq_table; + __be32 __iomem *doorbell; + u32 cons_index; + struct mlx5_frag_buf buf; + int size; + unsigned int irqn; + u8 eqn; + int nent; + struct list_head list; + struct mlx5_rsc_debug *dbg; + enum mlx5_eq_type type; + union { + struct mlx5_eq_tasklet tasklet_ctx; +#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING + struct mlx5_eq_pagefault pf_ctx; +#endif + }; +}; + +int mlx5_eq_table_init(struct mlx5_core_dev *dev); +void mlx5_eq_table_cleanup(struct mlx5_core_dev *dev); +int mlx5_eq_table_create(struct mlx5_core_dev *dev); +void mlx5_eq_table_destroy(struct mlx5_core_dev *dev); +int mlx5_eq_add_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq); +int mlx5_eq_del_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq); +struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn); +struct mlx5_eq *mlx5_get_async_eq(struct mlx5_core_dev *dev); +u32 mlx5_eq_poll_irq_disabled(struct mlx5_eq *eq); +void mlx5_cq_tasklet_cb(unsigned long data); +struct cpumask *mlx5_eq_comp_cpumask(struct mlx5_core_dev *dev, int ix); + +/* This function should only be called after mlx5_cmd_force_teardown_hca */ +void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev); + +#ifdef CONFIG_RFS_ACCEL +struct cpu_rmap *mlx5_eq_table_get_rmap(struct mlx5_core_dev *dev); +#endif + +#endif diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 21cc9bbc2563..5d11ef92c8b6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -54,6 +54,7 @@ #include #include "mlx5_core.h" #include "fs_core.h" +#include "lib/eq.h" #include "lib/mpfs.h" #include "eswitch.h" #include "lib/mlx5.h" @@ -728,7 +729,7 @@ static int mlx5_init_once(struct mlx5_core_dev *dev, struct mlx5_priv *priv) goto out; } - err = mlx5_eq_init(dev); + err = mlx5_eq_table_init(dev); if (err) { dev_err(&pdev->dev, "failed to initialize eq\n"); goto out; @@ -802,7 +803,7 @@ err_tables_cleanup: mlx5_cq_debugfs_cleanup(dev); err_eq_cleanup: - mlx5_eq_cleanup(dev); + mlx5_eq_table_cleanup(dev); out: return err; @@ -823,7 +824,7 @@ static void mlx5_cleanup_once(struct mlx5_core_dev *dev) mlx5_cleanup_srq_table(dev); mlx5_cleanup_qp_table(dev); mlx5_cq_debugfs_cleanup(dev); - mlx5_eq_cleanup(dev); + mlx5_eq_table_cleanup(dev); } static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index 3fa6d26875fe..4d39adcfb0eb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -124,21 +124,6 @@ int mlx5_destroy_scheduling_element_cmd(struct mlx5_core_dev *dev, u8 hierarchy, int mlx5_wait_for_vf_pages(struct mlx5_core_dev *dev); u64 mlx5_read_internal_timer(struct mlx5_core_dev *dev); -int mlx5_eq_init(struct mlx5_core_dev *dev); -void mlx5_eq_cleanup(struct mlx5_core_dev *dev); -int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, - int nent, u64 mask, const char *name, - enum mlx5_eq_type type); -int mlx5_destroy_unmap_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq); -int mlx5_eq_add_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq); -int mlx5_eq_del_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq); -int mlx5_eq_table_create(struct mlx5_core_dev *dev); -void mlx5_eq_table_destroy(struct mlx5_core_dev *dev); -/* This function should only be called after mlx5_cmd_force_teardown_hca */ -void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev); -struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn); -u32 mlx5_eq_poll_irq_disabled(struct mlx5_eq *eq); -void mlx5_cq_tasklet_cb(unsigned long data); void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool forced); int mlx5_debug_eq_add(struct mlx5_core_dev *dev, struct mlx5_eq *eq); void mlx5_debug_eq_remove(struct mlx5_core_dev *dev, struct mlx5_eq *eq); diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index dcc3f7aa8572..4d6246cb6c19 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -84,18 +84,6 @@ enum { MLX5_MAX_PORTS = 2, }; -enum { - MLX5_EQ_VEC_PAGES = 0, - MLX5_EQ_VEC_CMD = 1, - MLX5_EQ_VEC_ASYNC = 2, - MLX5_EQ_VEC_PFAULT = 3, - MLX5_EQ_VEC_COMP_BASE, -}; - -enum { - MLX5_MAX_IRQ_NAME = 32 -}; - enum { MLX5_ATOMIC_MODE_OFFSET = 16, MLX5_ATOMIC_MODE_IB_COMP = 1, @@ -366,49 +354,6 @@ struct mlx5_frag_buf_ctrl { u8 log_frag_strides; }; -struct mlx5_eq_tasklet { - struct list_head list; - struct list_head process_list; - struct tasklet_struct task; - /* lock on completion tasklet list */ - spinlock_t lock; -}; - -struct mlx5_eq_pagefault { - struct work_struct work; - /* Pagefaults lock */ - spinlock_t lock; - struct workqueue_struct *wq; - mempool_t *pool; -}; - -struct mlx5_cq_table { - /* protect radix tree */ - spinlock_t lock; - struct radix_tree_root tree; -}; - -struct mlx5_eq { - struct mlx5_core_dev *dev; - struct mlx5_cq_table cq_table; - __be32 __iomem *doorbell; - u32 cons_index; - struct mlx5_frag_buf buf; - int size; - unsigned int irqn; - u8 eqn; - int nent; - struct list_head list; - struct mlx5_rsc_debug *dbg; - enum mlx5_eq_type type; - union { - struct mlx5_eq_tasklet tasklet_ctx; -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - struct mlx5_eq_pagefault pf_ctx; -#endif - }; -}; - struct mlx5_core_psv { u32 psv_idx; struct psv_layout { @@ -475,21 +420,6 @@ struct mlx5_core_srq { u16 uid; }; -struct mlx5_eq_table { - struct list_head comp_eqs_list; - struct mlx5_eq pages_eq; - struct mlx5_eq async_eq; - struct mlx5_eq cmd_eq; -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - struct mlx5_eq pfault_eq; -#endif - int num_comp_vectors; - struct mlx5_irq_info *irq_info; -#ifdef CONFIG_RFS_ACCEL - struct cpu_rmap *rmap; -#endif -}; - struct mlx5_uars_page { void __iomem *map; bool wc; @@ -572,11 +502,6 @@ struct mlx5_core_sriov { int enabled_vfs; }; -struct mlx5_irq_info { - cpumask_var_t mask; - char name[MLX5_MAX_IRQ_NAME]; -}; - struct mlx5_fc_stats { spinlock_t counters_idr_lock; /* protects counters_idr */ struct idr counters_idr; @@ -594,6 +519,7 @@ struct mlx5_mpfs; struct mlx5_eswitch; struct mlx5_lag; struct mlx5_pagefault; +struct mlx5_eq_table; struct mlx5_rate_limit { u32 rate; @@ -643,7 +569,7 @@ struct mlx5_port_module_event_stats { struct mlx5_priv { char name[MLX5_MAX_NAME_LEN]; - struct mlx5_eq_table eq_table; + struct mlx5_eq_table *eq_table; /* pages stuff */ struct workqueue_struct *pg_wq; @@ -1148,6 +1074,9 @@ int mlx5_alloc_bfreg(struct mlx5_core_dev *mdev, struct mlx5_sq_bfreg *bfreg, bool map_wc, bool fast_path); void mlx5_free_bfreg(struct mlx5_core_dev *mdev, struct mlx5_sq_bfreg *bfreg); +unsigned int mlx5_comp_vectors_count(struct mlx5_core_dev *dev); +struct cpumask * +mlx5_comp_irq_get_affinity_mask(struct mlx5_core_dev *dev, int vector); unsigned int mlx5_core_reserved_gids_count(struct mlx5_core_dev *dev); int mlx5_core_roce_gid_set(struct mlx5_core_dev *dev, unsigned int index, u8 roce_version, u8 roce_l3_type, const u8 *gid, @@ -1299,10 +1228,4 @@ enum { MLX5_TRIGGERED_CMD_COMP = (u64)1 << 32, }; -static inline const struct cpumask * -mlx5_get_vector_affinity_hint(struct mlx5_core_dev *dev, int vector) -{ - return dev->priv.eq_table.irq_info[vector + MLX5_EQ_VEC_COMP_BASE].mask; -} - #endif /* MLX5_DRIVER_H */ -- cgit v1.2.3-55-g7522 From 16d760839ceef510cf95cbfadc069c4473c7a277 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 19 Nov 2018 10:52:39 -0800 Subject: net/mlx5: EQ, Different EQ types In mlx5 we have three types of usages for EQs, 1. Asynchronous EQs, used internally by mlx5 core for a. FW command completions b. FW page requests c. one EQ for all other Asynchronous events 2. Completion EQs, used for CQ completion (we create one per core) 3. *Special type of EQ (page fault) used for RDMA on demand paging (ODP). *The 3rd type shouldn't be special at least in mlx5 core, it is yet another async events EQ with specific use case, it will be removed in the next two patches, and will completely move its logic to mlx5_ib, as it is rdma specific. In this patch we remove use case (eq type) specific fields from struct mlx5_eq into a new eq type specific structures. struct mlx5_eq_async; truct mlx5_eq_comp; struct mlx5_eq_pagefault; Separate between their type specific flows. In the future we will allow users to create there own generic EQs. for now we will allow only one for ODP in next patches. We will introduce event listeners registration API for those who want to receive mlx5 async events. After that mlx5 eq handling will be clean from feature/user specific handling. Signed-off-by: Saeed Mahameed Reviewed-by: Leon Romanovsky Reviewed-by: Tariq Toukan Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/cq.c | 10 +- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 8 +- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 376 ++++++++++++--------- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/health.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h | 53 +-- drivers/net/ethernet/mellanox/mlx5/core/main.c | 2 +- .../net/ethernet/mellanox/mlx5/core/mlx5_core.h | 4 - include/linux/mlx5/cq.h | 2 +- include/linux/mlx5/driver.h | 10 +- 10 files changed, 270 insertions(+), 199 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cq.c b/drivers/net/ethernet/mellanox/mlx5/core/cq.c index 6e55d2f37c6d..713a17ee3751 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/cq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/cq.c @@ -93,10 +93,10 @@ int mlx5_core_create_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq, u32 dout[MLX5_ST_SZ_DW(destroy_cq_out)]; u32 out[MLX5_ST_SZ_DW(create_cq_out)]; u32 din[MLX5_ST_SZ_DW(destroy_cq_in)]; - struct mlx5_eq *eq; + struct mlx5_eq_comp *eq; int err; - eq = mlx5_eqn2eq(dev, eqn); + eq = mlx5_eqn2comp_eq(dev, eqn); if (IS_ERR(eq)) return PTR_ERR(eq); @@ -120,7 +120,7 @@ int mlx5_core_create_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq, INIT_LIST_HEAD(&cq->tasklet_ctx.list); /* Add to comp EQ CQ tree to recv comp events */ - err = mlx5_eq_add_cq(eq, cq); + err = mlx5_eq_add_cq(&eq->core, cq); if (err) goto err_cmd; @@ -140,7 +140,7 @@ int mlx5_core_create_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq, return 0; err_cq_add: - mlx5_eq_del_cq(eq, cq); + mlx5_eq_del_cq(&eq->core, cq); err_cmd: memset(din, 0, sizeof(din)); memset(dout, 0, sizeof(dout)); @@ -162,7 +162,7 @@ int mlx5_core_destroy_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq) if (err) return err; - err = mlx5_eq_del_cq(cq->eq, cq); + err = mlx5_eq_del_cq(&cq->eq->core, cq); if (err) return err; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index c23caade31bf..0d495a6b3949 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -320,7 +320,7 @@ static void mlx5e_enable_async_events(struct mlx5e_priv *priv) static void mlx5e_disable_async_events(struct mlx5e_priv *priv) { clear_bit(MLX5E_STATE_ASYNC_EVENTS_ENABLED, &priv->state); - synchronize_irq(pci_irq_vector(priv->mdev->pdev, MLX5_EQ_VEC_ASYNC)); + mlx5_eq_synchronize_async_irq(priv->mdev); } static inline void mlx5e_build_umr_wqe(struct mlx5e_rq *rq, @@ -4117,17 +4117,17 @@ static netdev_features_t mlx5e_features_check(struct sk_buff *skb, static bool mlx5e_tx_timeout_eq_recover(struct net_device *dev, struct mlx5e_txqsq *sq) { - struct mlx5_eq *eq = sq->cq.mcq.eq; + struct mlx5_eq_comp *eq = sq->cq.mcq.eq; u32 eqe_count; netdev_err(dev, "EQ 0x%x: Cons = 0x%x, irqn = 0x%x\n", - eq->eqn, eq->cons_index, eq->irqn); + eq->core.eqn, eq->core.cons_index, eq->core.irqn); eqe_count = mlx5_eq_poll_irq_disabled(eq); if (!eqe_count) return false; - netdev_err(dev, "Recover %d eqes on EQ 0x%x\n", eqe_count, eq->eqn); + netdev_err(dev, "Recover %d eqes on EQ 0x%x\n", eqe_count, eq->core.eqn); sq->channel->stats->eq_rearm++; return true; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 32ce20221c44..252c9f0569b1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -72,13 +72,16 @@ struct mlx5_irq_info { }; struct mlx5_eq_table { - struct list_head comp_eqs_list; - struct mlx5_eq pages_eq; - struct mlx5_eq async_eq; - struct mlx5_eq cmd_eq; + struct list_head comp_eqs_list; + struct mlx5_eq pages_eq; + struct mlx5_eq async_eq; + struct mlx5_eq cmd_eq; + #ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - struct mlx5_eq pfault_eq; + struct mlx5_eq_pagefault pfault_eq; #endif + struct mutex lock; /* sync async eqs creations */ + u8 num_async_eqs; int num_comp_vectors; struct mlx5_irq_info *irq_info; #ifdef CONFIG_RFS_ACCEL @@ -224,24 +227,24 @@ static void eqe_pf_action(struct work_struct *work) struct mlx5_pagefault *pfault = container_of(work, struct mlx5_pagefault, work); - struct mlx5_eq *eq = pfault->eq; + struct mlx5_eq_pagefault *eq = pfault->eq; - mlx5_core_page_fault(eq->dev, pfault); - mempool_free(pfault, eq->pf_ctx.pool); + mlx5_core_page_fault(eq->core.dev, pfault); + mempool_free(pfault, eq->pool); } -static void eq_pf_process(struct mlx5_eq *eq) +static void eq_pf_process(struct mlx5_eq_pagefault *eq) { - struct mlx5_core_dev *dev = eq->dev; + struct mlx5_core_dev *dev = eq->core.dev; struct mlx5_eqe_page_fault *pf_eqe; struct mlx5_pagefault *pfault; struct mlx5_eqe *eqe; int set_ci = 0; - while ((eqe = next_eqe_sw(eq))) { - pfault = mempool_alloc(eq->pf_ctx.pool, GFP_ATOMIC); + while ((eqe = next_eqe_sw(&eq->core))) { + pfault = mempool_alloc(eq->pool, GFP_ATOMIC); if (!pfault) { - schedule_work(&eq->pf_ctx.work); + schedule_work(&eq->work); break; } @@ -311,30 +314,30 @@ static void eq_pf_process(struct mlx5_eq *eq) pfault->eq = eq; INIT_WORK(&pfault->work, eqe_pf_action); - queue_work(eq->pf_ctx.wq, &pfault->work); + queue_work(eq->wq, &pfault->work); - ++eq->cons_index; + ++eq->core.cons_index; ++set_ci; if (unlikely(set_ci >= MLX5_NUM_SPARE_EQE)) { - eq_update_ci(eq, 0); + eq_update_ci(&eq->core, 0); set_ci = 0; } } - eq_update_ci(eq, 1); + eq_update_ci(&eq->core, 1); } static irqreturn_t mlx5_eq_pf_int(int irq, void *eq_ptr) { - struct mlx5_eq *eq = eq_ptr; + struct mlx5_eq_pagefault *eq = eq_ptr; unsigned long flags; - if (spin_trylock_irqsave(&eq->pf_ctx.lock, flags)) { + if (spin_trylock_irqsave(&eq->lock, flags)) { eq_pf_process(eq); - spin_unlock_irqrestore(&eq->pf_ctx.lock, flags); + spin_unlock_irqrestore(&eq->lock, flags); } else { - schedule_work(&eq->pf_ctx.work); + schedule_work(&eq->work); } return IRQ_HANDLED; @@ -352,35 +355,61 @@ static void mempool_refill(mempool_t *pool) static void eq_pf_action(struct work_struct *work) { - struct mlx5_eq *eq = container_of(work, struct mlx5_eq, pf_ctx.work); + struct mlx5_eq_pagefault *eq = + container_of(work, struct mlx5_eq_pagefault, work); - mempool_refill(eq->pf_ctx.pool); + mempool_refill(eq->pool); - spin_lock_irq(&eq->pf_ctx.lock); + spin_lock_irq(&eq->lock); eq_pf_process(eq); - spin_unlock_irq(&eq->pf_ctx.lock); + spin_unlock_irq(&eq->lock); } -static int init_pf_ctx(struct mlx5_eq_pagefault *pf_ctx, const char *name) +static int +create_pf_eq(struct mlx5_core_dev *dev, struct mlx5_eq_pagefault *eq) { - spin_lock_init(&pf_ctx->lock); - INIT_WORK(&pf_ctx->work, eq_pf_action); + int err; - pf_ctx->wq = alloc_workqueue(name, - WQ_HIGHPRI | WQ_UNBOUND | WQ_MEM_RECLAIM, - MLX5_NUM_CMD_EQE); - if (!pf_ctx->wq) + spin_lock_init(&eq->lock); + INIT_WORK(&eq->work, eq_pf_action); + + eq->pool = mempool_create_kmalloc_pool(MLX5_NUM_PF_DRAIN, + sizeof(struct mlx5_pagefault)); + if (!eq->pool) return -ENOMEM; - pf_ctx->pool = mempool_create_kmalloc_pool - (MLX5_NUM_PF_DRAIN, sizeof(struct mlx5_pagefault)); - if (!pf_ctx->pool) + eq->wq = alloc_workqueue("mlx5_page_fault", + WQ_HIGHPRI | WQ_UNBOUND | WQ_MEM_RECLAIM, + MLX5_NUM_CMD_EQE); + if (!eq->wq) { + err = -ENOMEM; + goto err_mempool; + } + + err = mlx5_create_async_eq(dev, &eq->core, MLX5_NUM_ASYNC_EQE, + 1 << MLX5_EVENT_TYPE_PAGE_FAULT, + "mlx5_page_fault_eq", mlx5_eq_pf_int); + if (err) goto err_wq; return 0; err_wq: - destroy_workqueue(pf_ctx->wq); - return -ENOMEM; + destroy_workqueue(eq->wq); +err_mempool: + mempool_destroy(eq->pool); + return err; +} + +static int destroy_pf_eq(struct mlx5_core_dev *dev, struct mlx5_eq_pagefault *eq) +{ + int err; + + err = mlx5_destroy_async_eq(dev, &eq->core); + cancel_work_sync(&eq->work); + destroy_workqueue(eq->wq); + mempool_destroy(eq->pool); + + return err; } int mlx5_core_page_fault_resume(struct mlx5_core_dev *dev, u32 token, @@ -444,37 +473,88 @@ static struct mlx5_core_cq *mlx5_eq_cq_get(struct mlx5_eq *eq, u32 cqn) return cq; } -static void mlx5_eq_cq_completion(struct mlx5_eq *eq, u32 cqn) +static void mlx5_eq_cq_event(struct mlx5_eq *eq, u32 cqn, int event_type) { struct mlx5_core_cq *cq = mlx5_eq_cq_get(eq, cqn); if (unlikely(!cq)) { - mlx5_core_warn(eq->dev, "Completion event for bogus CQ 0x%x\n", cqn); + mlx5_core_warn(eq->dev, "Async event for bogus CQ 0x%x\n", cqn); return; } - ++cq->arm_sn; - - cq->comp(cq); + cq->event(cq, event_type); mlx5_cq_put(cq); } -static void mlx5_eq_cq_event(struct mlx5_eq *eq, u32 cqn, int event_type) +static irqreturn_t mlx5_eq_comp_int(int irq, void *eq_ptr) { - struct mlx5_core_cq *cq = mlx5_eq_cq_get(eq, cqn); + struct mlx5_eq_comp *eq_comp = eq_ptr; + struct mlx5_eq *eq = eq_ptr; + struct mlx5_eqe *eqe; + int set_ci = 0; + u32 cqn = -1; - if (unlikely(!cq)) { - mlx5_core_warn(eq->dev, "Async event for bogus CQ 0x%x\n", cqn); - return; + while ((eqe = next_eqe_sw(eq))) { + struct mlx5_core_cq *cq; + /* Make sure we read EQ entry contents after we've + * checked the ownership bit. + */ + dma_rmb(); + /* Assume (eqe->type) is always MLX5_EVENT_TYPE_COMP */ + cqn = be32_to_cpu(eqe->data.comp.cqn) & 0xffffff; + + cq = mlx5_eq_cq_get(eq, cqn); + if (likely(cq)) { + ++cq->arm_sn; + cq->comp(cq); + mlx5_cq_put(cq); + } else { + mlx5_core_warn(eq->dev, "Completion event for bogus CQ 0x%x\n", cqn); + } + + ++eq->cons_index; + ++set_ci; + + /* The HCA will think the queue has overflowed if we + * don't tell it we've been processing events. We + * create our EQs with MLX5_NUM_SPARE_EQE extra + * entries, so we must update our consumer index at + * least that often. + */ + if (unlikely(set_ci >= MLX5_NUM_SPARE_EQE)) { + eq_update_ci(eq, 0); + set_ci = 0; + } } - cq->event(cq, event_type); + eq_update_ci(eq, 1); - mlx5_cq_put(cq); + if (cqn != -1) + tasklet_schedule(&eq_comp->tasklet_ctx.task); + + return IRQ_HANDLED; } -static irqreturn_t mlx5_eq_int(int irq, void *eq_ptr) +/* Some architectures don't latch interrupts when they are disabled, so using + * mlx5_eq_poll_irq_disabled could end up losing interrupts while trying to + * avoid losing them. It is not recommended to use it, unless this is the last + * resort. + */ +u32 mlx5_eq_poll_irq_disabled(struct mlx5_eq_comp *eq) +{ + u32 count_eqe; + + disable_irq(eq->core.irqn); + count_eqe = eq->core.cons_index; + mlx5_eq_comp_int(eq->core.irqn, eq); + count_eqe = eq->core.cons_index - count_eqe; + enable_irq(eq->core.irqn); + + return count_eqe; +} + +static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) { struct mlx5_eq *eq = eq_ptr; struct mlx5_core_dev *dev = eq->dev; @@ -494,10 +574,6 @@ static irqreturn_t mlx5_eq_int(int irq, void *eq_ptr) mlx5_core_dbg(eq->dev, "eqn %d, eqe type %s\n", eq->eqn, eqe_type_str(eqe->type)); switch (eqe->type) { - case MLX5_EVENT_TYPE_COMP: - cqn = be32_to_cpu(eqe->data.comp.cqn) & 0xffffff; - mlx5_eq_cq_completion(eq, cqn); - break; case MLX5_EVENT_TYPE_DCT_DRAINED: rsn = be32_to_cpu(eqe->data.dct.dctn) & 0xffffff; rsn |= (MLX5_RES_DCT << MLX5_USER_INDEX_LEN); @@ -619,30 +695,9 @@ static irqreturn_t mlx5_eq_int(int irq, void *eq_ptr) eq_update_ci(eq, 1); - if (cqn != -1) - tasklet_schedule(&eq->tasklet_ctx.task); - return IRQ_HANDLED; } -/* Some architectures don't latch interrupts when they are disabled, so using - * mlx5_eq_poll_irq_disabled could end up losing interrupts while trying to - * avoid losing them. It is not recommended to use it, unless this is the last - * resort. - */ -u32 mlx5_eq_poll_irq_disabled(struct mlx5_eq *eq) -{ - u32 count_eqe; - - disable_irq(eq->irqn); - count_eqe = eq->cons_index; - mlx5_eq_int(eq->irqn, eq); - count_eqe = eq->cons_index - count_eqe; - enable_irq(eq->irqn); - - return count_eqe; -} - static void init_eq_buf(struct mlx5_eq *eq) { struct mlx5_eqe *eqe; @@ -656,13 +711,12 @@ static void init_eq_buf(struct mlx5_eq *eq) static int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, - int nent, u64 mask, const char *name, enum mlx5_eq_type type) + int nent, u64 mask, const char *name, irq_handler_t handler) { struct mlx5_eq_table *eq_table = dev->priv.eq_table; struct mlx5_cq_table *cq_table = &eq->cq_table; u32 out[MLX5_ST_SZ_DW(create_eq_out)] = {0}; struct mlx5_priv *priv = &dev->priv; - irq_handler_t handler; __be64 *pas; void *eqc; int inlen; @@ -674,20 +728,12 @@ mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, spin_lock_init(&cq_table->lock); INIT_RADIX_TREE(&cq_table->tree, GFP_ATOMIC); - eq->type = type; eq->nent = roundup_pow_of_two(nent + MLX5_NUM_SPARE_EQE); eq->cons_index = 0; err = mlx5_buf_alloc(dev, eq->nent * MLX5_EQE_SIZE, &eq->buf); if (err) return err; -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - if (type == MLX5_EQ_TYPE_PF) - handler = mlx5_eq_pf_int; - else -#endif - handler = mlx5_eq_int; - init_eq_buf(eq); inlen = MLX5_ST_SZ_BYTES(create_eq_in) + @@ -732,21 +778,6 @@ mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, if (err) goto err_irq; -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - if (type == MLX5_EQ_TYPE_PF) { - err = init_pf_ctx(&eq->pf_ctx, name); - if (err) - goto err_irq; - } else -#endif - { - INIT_LIST_HEAD(&eq->tasklet_ctx.list); - INIT_LIST_HEAD(&eq->tasklet_ctx.process_list); - spin_lock_init(&eq->tasklet_ctx.lock); - tasklet_init(&eq->tasklet_ctx.task, mlx5_cq_tasklet_cb, - (unsigned long)&eq->tasklet_ctx); - } - /* EQs are created in ARMED state */ eq_update_ci(eq, 1); @@ -780,15 +811,6 @@ static int mlx5_destroy_unmap_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq) eq->eqn); synchronize_irq(eq->irqn); - if (eq->type == MLX5_EQ_TYPE_COMP) { - tasklet_disable(&eq->tasklet_ctx.task); -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - } else if (eq->type == MLX5_EQ_TYPE_PF) { - cancel_work_sync(&eq->pf_ctx.work); - destroy_workqueue(eq->pf_ctx.wq); - mempool_destroy(eq->pf_ctx.pool); -#endif - } mlx5_buf_free(dev, &eq->buf); return err; @@ -841,8 +863,15 @@ int mlx5_eq_table_init(struct mlx5_core_dev *dev) err = mlx5_eq_debugfs_init(dev); if (err) - kvfree(eq_table); + goto kvfree_eq_table; + mutex_init(&eq_table->lock); + + return 0; + +kvfree_eq_table: + kvfree(eq_table); + dev->priv.eq_table = NULL; return err; } @@ -854,6 +883,43 @@ void mlx5_eq_table_cleanup(struct mlx5_core_dev *dev) /* Async EQs */ +int mlx5_create_async_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, + int nent, u64 mask, const char *name, irq_handler_t handler) +{ + struct mlx5_eq_table *eq_table = dev->priv.eq_table; + u8 vecdix; + int err; + + mutex_lock(&eq_table->lock); + if (eq_table->num_async_eqs >= MLX5_EQ_MAX_ASYNC_EQS) { + err = -ENOSPC; + goto unlock; + } + + vecdix = eq_table->num_async_eqs + 1; + + err = mlx5_create_map_eq(dev, eq, vecdix, nent, mask, name, handler); + if (!err) + eq_table->num_async_eqs++; + +unlock: + mutex_unlock(&eq_table->lock); + return err; +} + +int mlx5_destroy_async_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq) +{ + struct mlx5_eq_table *eq_table = dev->priv.eq_table; + int err; + + mutex_lock(&eq_table->lock); + err = mlx5_destroy_unmap_eq(dev, eq); + if (!err) + eq_table->num_async_eqs--; + mutex_unlock(&eq_table->lock); + return err; +} + static int create_async_eqs(struct mlx5_core_dev *dev) { struct mlx5_eq_table *table = dev->priv.eq_table; @@ -887,9 +953,9 @@ static int create_async_eqs(struct mlx5_core_dev *dev) if (MLX5_CAP_MCAM_REG(dev, tracer_registers)) async_event_mask |= (1ull << MLX5_EVENT_TYPE_DEVICE_TRACER); - err = mlx5_create_map_eq(dev, &table->cmd_eq, MLX5_EQ_VEC_CMD, - MLX5_NUM_CMD_EQE, 1ull << MLX5_EVENT_TYPE_CMD, - "mlx5_cmd_eq", MLX5_EQ_TYPE_ASYNC); + err = mlx5_create_async_eq(dev, &table->cmd_eq, MLX5_NUM_CMD_EQE, + 1ull << MLX5_EVENT_TYPE_CMD, "mlx5_cmd_eq", + mlx5_eq_async_int); if (err) { mlx5_core_warn(dev, "failed to create cmd EQ %d\n", err); return err; @@ -897,19 +963,15 @@ static int create_async_eqs(struct mlx5_core_dev *dev) mlx5_cmd_use_events(dev); - err = mlx5_create_map_eq(dev, &table->async_eq, MLX5_EQ_VEC_ASYNC, - MLX5_NUM_ASYNC_EQE, async_event_mask, - "mlx5_async_eq", MLX5_EQ_TYPE_ASYNC); + err = mlx5_create_async_eq(dev, &table->async_eq, MLX5_NUM_ASYNC_EQE, + async_event_mask, "mlx5_async_eq", mlx5_eq_async_int); if (err) { mlx5_core_warn(dev, "failed to create async EQ %d\n", err); goto err1; } - err = mlx5_create_map_eq(dev, &table->pages_eq, - MLX5_EQ_VEC_PAGES, - /* TODO: sriov max_vf + */ 1, - 1 << MLX5_EVENT_TYPE_PAGE_REQUEST, "mlx5_pages_eq", - MLX5_EQ_TYPE_ASYNC); + err = mlx5_create_async_eq(dev, &table->pages_eq, /* TODO: sriov max_vf + */ 1, + 1 << MLX5_EVENT_TYPE_PAGE_REQUEST, "mlx5_pages_eq", mlx5_eq_async_int); if (err) { mlx5_core_warn(dev, "failed to create pages EQ %d\n", err); goto err2; @@ -917,12 +979,7 @@ static int create_async_eqs(struct mlx5_core_dev *dev) #ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING if (MLX5_CAP_GEN(dev, pg)) { - err = mlx5_create_map_eq(dev, &table->pfault_eq, - MLX5_EQ_VEC_PFAULT, - MLX5_NUM_ASYNC_EQE, - 1 << MLX5_EVENT_TYPE_PAGE_FAULT, - "mlx5_page_fault_eq", - MLX5_EQ_TYPE_PF); + err = create_pf_eq(dev, &table->pfault_eq); if (err) { mlx5_core_warn(dev, "failed to create page fault EQ %d\n", err); @@ -932,17 +989,17 @@ static int create_async_eqs(struct mlx5_core_dev *dev) return err; err3: - mlx5_destroy_unmap_eq(dev, &table->pages_eq); + mlx5_destroy_async_eq(dev, &table->pages_eq); #else return err; #endif err2: - mlx5_destroy_unmap_eq(dev, &table->async_eq); + mlx5_destroy_async_eq(dev, &table->async_eq); err1: mlx5_cmd_use_polling(dev); - mlx5_destroy_unmap_eq(dev, &table->cmd_eq); + mlx5_destroy_async_eq(dev, &table->cmd_eq); return err; } @@ -953,25 +1010,25 @@ static void destroy_async_eqs(struct mlx5_core_dev *dev) #ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING if (MLX5_CAP_GEN(dev, pg)) { - err = mlx5_destroy_unmap_eq(dev, &table->pfault_eq); + err = destroy_pf_eq(dev, &table->pfault_eq); if (err) mlx5_core_err(dev, "failed to destroy page fault eq, err(%d)\n", err); } #endif - err = mlx5_destroy_unmap_eq(dev, &table->pages_eq); + err = mlx5_destroy_async_eq(dev, &table->pages_eq); if (err) mlx5_core_err(dev, "failed to destroy pages eq, err(%d)\n", err); - err = mlx5_destroy_unmap_eq(dev, &table->async_eq); + err = mlx5_destroy_async_eq(dev, &table->async_eq); if (err) mlx5_core_err(dev, "failed to destroy async eq, err(%d)\n", err); mlx5_cmd_use_polling(dev); - err = mlx5_destroy_unmap_eq(dev, &table->cmd_eq); + err = mlx5_destroy_async_eq(dev, &table->cmd_eq); if (err) mlx5_core_err(dev, "failed to destroy command eq, err(%d)\n", err); @@ -982,6 +1039,16 @@ struct mlx5_eq *mlx5_get_async_eq(struct mlx5_core_dev *dev) return &dev->priv.eq_table->async_eq; } +void mlx5_eq_synchronize_async_irq(struct mlx5_core_dev *dev) +{ + synchronize_irq(dev->priv.eq_table->async_eq.irqn); +} + +void mlx5_eq_synchronize_cmd_irq(struct mlx5_core_dev *dev) +{ + synchronize_irq(dev->priv.eq_table->cmd_eq.irqn); +} + /* Completion EQs */ static int set_comp_irq_affinity_hint(struct mlx5_core_dev *mdev, int i) @@ -1048,7 +1115,7 @@ static void clear_comp_irqs_affinity_hints(struct mlx5_core_dev *mdev) static void destroy_comp_eqs(struct mlx5_core_dev *dev) { struct mlx5_eq_table *table = dev->priv.eq_table; - struct mlx5_eq *eq, *n; + struct mlx5_eq_comp *eq, *n; clear_comp_irqs_affinity_hints(dev); @@ -1060,9 +1127,10 @@ static void destroy_comp_eqs(struct mlx5_core_dev *dev) #endif list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) { list_del(&eq->list); - if (mlx5_destroy_unmap_eq(dev, eq)) - mlx5_core_warn(dev, "failed to destroy EQ 0x%x\n", - eq->eqn); + if (mlx5_destroy_unmap_eq(dev, &eq->core)) + mlx5_core_warn(dev, "failed to destroy comp EQ 0x%x\n", + eq->core.eqn); + tasklet_disable(&eq->tasklet_ctx.task); kfree(eq); } } @@ -1071,7 +1139,7 @@ static int create_comp_eqs(struct mlx5_core_dev *dev) { struct mlx5_eq_table *table = dev->priv.eq_table; char name[MLX5_MAX_IRQ_NAME]; - struct mlx5_eq *eq; + struct mlx5_eq_comp *eq; int ncomp_vec; int nent; int err; @@ -1094,17 +1162,23 @@ static int create_comp_eqs(struct mlx5_core_dev *dev) goto clean; } + INIT_LIST_HEAD(&eq->tasklet_ctx.list); + INIT_LIST_HEAD(&eq->tasklet_ctx.process_list); + spin_lock_init(&eq->tasklet_ctx.lock); + tasklet_init(&eq->tasklet_ctx.task, mlx5_cq_tasklet_cb, + (unsigned long)&eq->tasklet_ctx); + #ifdef CONFIG_RFS_ACCEL irq_cpu_rmap_add(table->rmap, pci_irq_vector(dev->pdev, vecidx)); #endif snprintf(name, MLX5_MAX_IRQ_NAME, "mlx5_comp%d", i); - err = mlx5_create_map_eq(dev, eq, vecidx, nent, 0, - name, MLX5_EQ_TYPE_COMP); + err = mlx5_create_map_eq(dev, &eq->core, vecidx, nent, 0, + name, mlx5_eq_comp_int); if (err) { kfree(eq); goto clean; } - mlx5_core_dbg(dev, "allocated completion EQN %d\n", eq->eqn); + mlx5_core_dbg(dev, "allocated completion EQN %d\n", eq->core.eqn); /* add tail, to keep the list ordered, for mlx5_vector2eqn to work */ list_add_tail(&eq->list, &table->comp_eqs_list); } @@ -1126,14 +1200,14 @@ int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn, unsigned int *irqn) { struct mlx5_eq_table *table = dev->priv.eq_table; - struct mlx5_eq *eq, *n; + struct mlx5_eq_comp *eq, *n; int err = -ENOENT; int i = 0; list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) { if (i++ == vector) { - *eqn = eq->eqn; - *irqn = eq->irqn; + *eqn = eq->core.eqn; + *irqn = eq->core.irqn; err = 0; break; } @@ -1166,13 +1240,13 @@ struct cpu_rmap *mlx5_eq_table_get_rmap(struct mlx5_core_dev *dev) #endif } -struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn) +struct mlx5_eq_comp *mlx5_eqn2comp_eq(struct mlx5_core_dev *dev, int eqn) { struct mlx5_eq_table *table = dev->priv.eq_table; - struct mlx5_eq *eq; + struct mlx5_eq_comp *eq; list_for_each_entry(eq, &table->comp_eqs_list, list) { - if (eq->eqn == eqn) + if (eq->core.eqn == eqn) return eq; } @@ -1183,7 +1257,7 @@ struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn) void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev) { struct mlx5_eq_table *table = dev->priv.eq_table; - struct mlx5_eq *eq; + struct mlx5_eq_comp *eq; clear_comp_irqs_affinity_hints(dev); @@ -1194,14 +1268,14 @@ void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev) } #endif list_for_each_entry(eq, &table->comp_eqs_list, list) - free_irq(eq->irqn, eq); + free_irq(eq->core.irqn, eq); free_irq(table->pages_eq.irqn, &table->pages_eq); free_irq(table->async_eq.irqn, &table->async_eq); free_irq(table->cmd_eq.irqn, &table->cmd_eq); #ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING if (MLX5_CAP_GEN(dev, pg)) - free_irq(table->pfault_eq.irqn, &table->pfault_eq); + free_irq(table->pfault_eq.core.irqn, &table->pfault_eq.core); #endif pci_free_irq_vectors(dev->pdev); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index 324606227b1a..2346b6ba3d54 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -1568,7 +1568,7 @@ static void esw_disable_vport(struct mlx5_eswitch *esw, int vport_num) /* Mark this vport as disabled to discard new events */ vport->enabled = false; - synchronize_irq(pci_irq_vector(esw->dev->pdev, MLX5_EQ_VEC_ASYNC)); + mlx5_eq_synchronize_async_irq(esw->dev); /* Wait for current already scheduled events to complete */ flush_workqueue(esw->work_queue); /* Disable events from this vport */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c index b5be6f0b9ed5..066883003aea 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/health.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c @@ -85,7 +85,7 @@ static void trigger_cmd_completions(struct mlx5_core_dev *dev) u64 vector; /* wait for pending handlers to complete */ - synchronize_irq(pci_irq_vector(dev->pdev, MLX5_EQ_VEC_CMD)); + mlx5_eq_synchronize_cmd_irq(dev); spin_lock_irqsave(&dev->cmd.alloc_lock, flags); vector = ~dev->cmd.bitmask & ((1ul << (1 << dev->cmd.log_sz)) - 1); if (!vector) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h index 48ee37797b3f..706d58383dbd 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h @@ -8,11 +8,8 @@ #define MLX5_MAX_IRQ_NAME (32) enum { - MLX5_EQ_VEC_PAGES = 0, - MLX5_EQ_VEC_CMD = 1, - MLX5_EQ_VEC_ASYNC = 2, - MLX5_EQ_VEC_PFAULT = 3, - MLX5_EQ_VEC_COMP_BASE, + MLX5_EQ_MAX_ASYNC_EQS = 4, /* mlx5_core needs at least 3 */ + MLX5_EQ_VEC_COMP_BASE = MLX5_EQ_MAX_ASYNC_EQS, }; struct mlx5_eq_tasklet { @@ -22,13 +19,6 @@ struct mlx5_eq_tasklet { spinlock_t lock; /* lock completion tasklet list */ }; -struct mlx5_eq_pagefault { - struct work_struct work; - spinlock_t lock; /* Pagefaults spinlock */ - struct workqueue_struct *wq; - mempool_t *pool; -}; - struct mlx5_cq_table { spinlock_t lock; /* protect radix tree */ struct radix_tree_root tree; @@ -44,29 +34,48 @@ struct mlx5_eq { unsigned int irqn; u8 eqn; int nent; - struct list_head list; struct mlx5_rsc_debug *dbg; - enum mlx5_eq_type type; - union { - struct mlx5_eq_tasklet tasklet_ctx; -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - struct mlx5_eq_pagefault pf_ctx; -#endif - }; +}; + +struct mlx5_eq_comp { + struct mlx5_eq core; /* Must be first */ + struct mlx5_eq_tasklet tasklet_ctx; + struct list_head list; +}; + +struct mlx5_eq_pagefault { + struct mlx5_eq core; /* Must be first */ + struct work_struct work; + spinlock_t lock; /* Pagefaults spinlock */ + struct workqueue_struct *wq; + mempool_t *pool; }; int mlx5_eq_table_init(struct mlx5_core_dev *dev); void mlx5_eq_table_cleanup(struct mlx5_core_dev *dev); int mlx5_eq_table_create(struct mlx5_core_dev *dev); void mlx5_eq_table_destroy(struct mlx5_core_dev *dev); +int mlx5_create_async_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, + int nent, u64 mask, const char *name, + irq_handler_t handler); +int mlx5_destroy_async_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq); + int mlx5_eq_add_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq); int mlx5_eq_del_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq); -struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn); +struct mlx5_eq_comp *mlx5_eqn2comp_eq(struct mlx5_core_dev *dev, int eqn); struct mlx5_eq *mlx5_get_async_eq(struct mlx5_core_dev *dev); -u32 mlx5_eq_poll_irq_disabled(struct mlx5_eq *eq); void mlx5_cq_tasklet_cb(unsigned long data); struct cpumask *mlx5_eq_comp_cpumask(struct mlx5_core_dev *dev, int ix); +u32 mlx5_eq_poll_irq_disabled(struct mlx5_eq_comp *eq); +void mlx5_eq_synchronize_async_irq(struct mlx5_core_dev *dev); +void mlx5_eq_synchronize_cmd_irq(struct mlx5_core_dev *dev); + +int mlx5_debug_eq_add(struct mlx5_core_dev *dev, struct mlx5_eq *eq); +void mlx5_debug_eq_remove(struct mlx5_core_dev *dev, struct mlx5_eq *eq); +int mlx5_eq_debugfs_init(struct mlx5_core_dev *dev); +void mlx5_eq_debugfs_cleanup(struct mlx5_core_dev *dev); + /* This function should only be called after mlx5_cmd_force_teardown_hca */ void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 5d11ef92c8b6..3de83fe65f2b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -53,8 +53,8 @@ #endif #include #include "mlx5_core.h" -#include "fs_core.h" #include "lib/eq.h" +#include "fs_core.h" #include "lib/mpfs.h" #include "eswitch.h" #include "lib/mlx5.h" diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index 4d39adcfb0eb..4728b027cb9e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -125,10 +125,6 @@ int mlx5_wait_for_vf_pages(struct mlx5_core_dev *dev); u64 mlx5_read_internal_timer(struct mlx5_core_dev *dev); void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool forced); -int mlx5_debug_eq_add(struct mlx5_core_dev *dev, struct mlx5_eq *eq); -void mlx5_debug_eq_remove(struct mlx5_core_dev *dev, struct mlx5_eq *eq); -int mlx5_eq_debugfs_init(struct mlx5_core_dev *dev); -void mlx5_eq_debugfs_cleanup(struct mlx5_core_dev *dev); int mlx5_cq_debugfs_init(struct mlx5_core_dev *dev); void mlx5_cq_debugfs_cleanup(struct mlx5_core_dev *dev); diff --git a/include/linux/mlx5/cq.h b/include/linux/mlx5/cq.h index 31a750570c38..28b757a64029 100644 --- a/include/linux/mlx5/cq.h +++ b/include/linux/mlx5/cq.h @@ -60,7 +60,7 @@ struct mlx5_core_cq { } tasklet_ctx; int reset_notify_added; struct list_head reset_notify; - struct mlx5_eq *eq; + struct mlx5_eq_comp *eq; u16 uid; }; diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 4d6246cb6c19..fe9b552aa649 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -210,14 +210,6 @@ enum mlx5_port_status { MLX5_PORT_DOWN = 2, }; -enum mlx5_eq_type { - MLX5_EQ_TYPE_COMP, - MLX5_EQ_TYPE_ASYNC, -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - MLX5_EQ_TYPE_PF, -#endif -}; - struct mlx5_bfreg_info { u32 *sys_pages; int num_low_latency_bfregs; @@ -692,7 +684,7 @@ struct mlx5_pagefault { } rdma; }; - struct mlx5_eq *eq; + struct mlx5_eq_pagefault *eq; struct work_struct work; }; -- cgit v1.2.3-55-g7522 From 7701707cb94ed4d1e63ae4fa5ef62a2345ef9db7 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 19 Nov 2018 10:52:40 -0800 Subject: net/mlx5: EQ, Generic EQ Add mlx5_eq_{create/destroy}_generic APIs and EQE access methods, for mlx5 core consumers generic EQs. This API will be used in downstream patch to move page fault (RDMA ODP) EQ logic into mlx5_ib rdma driver, hence it will use a generic EQ. Current mlx5 EQ allocation scheme: On load mlx5 allocates 4 (for async) + #cores (for data completions) MSIX vectors, mlx5 core will assign 3 MSIX vectors for internal async EQs and will use all of the #cores MSIX vectors for completion EQs, (One vector is going to be reserved for a generic EQ). After this patch an external user (e.g mlx5_ib) of mlx5_core can use this new API to create new generic EQs with the reserved msix vector index for that eq. Signed-off-by: Saeed Mahameed Reviewed-by: Leon Romanovsky Reviewed-by: Tariq Toukan Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 243 +++++++++++++++++------ drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h | 12 +- include/linux/mlx5/eq.h | 39 ++++ 3 files changed, 221 insertions(+), 73 deletions(-) create mode 100644 include/linux/mlx5/eq.h (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 252c9f0569b1..ec1f5018546e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -33,6 +33,7 @@ #include #include #include +#include #include #ifdef CONFIG_RFS_ACCEL #include @@ -69,6 +70,7 @@ enum { struct mlx5_irq_info { cpumask_var_t mask; char name[MLX5_MAX_IRQ_NAME]; + void *context; /* dev_id provided to request_irq */ }; struct mlx5_eq_table { @@ -81,7 +83,6 @@ struct mlx5_eq_table { struct mlx5_eq_pagefault pfault_eq; #endif struct mutex lock; /* sync async eqs creations */ - u8 num_async_eqs; int num_comp_vectors; struct mlx5_irq_info *irq_info; #ifdef CONFIG_RFS_ACCEL @@ -229,19 +230,19 @@ static void eqe_pf_action(struct work_struct *work) work); struct mlx5_eq_pagefault *eq = pfault->eq; - mlx5_core_page_fault(eq->core.dev, pfault); + mlx5_core_page_fault(eq->core->dev, pfault); mempool_free(pfault, eq->pool); } static void eq_pf_process(struct mlx5_eq_pagefault *eq) { - struct mlx5_core_dev *dev = eq->core.dev; + struct mlx5_core_dev *dev = eq->core->dev; struct mlx5_eqe_page_fault *pf_eqe; struct mlx5_pagefault *pfault; struct mlx5_eqe *eqe; int set_ci = 0; - while ((eqe = next_eqe_sw(&eq->core))) { + while ((eqe = next_eqe_sw(eq->core))) { pfault = mempool_alloc(eq->pool, GFP_ATOMIC); if (!pfault) { schedule_work(&eq->work); @@ -316,16 +317,16 @@ static void eq_pf_process(struct mlx5_eq_pagefault *eq) INIT_WORK(&pfault->work, eqe_pf_action); queue_work(eq->wq, &pfault->work); - ++eq->core.cons_index; + ++eq->core->cons_index; ++set_ci; if (unlikely(set_ci >= MLX5_NUM_SPARE_EQE)) { - eq_update_ci(&eq->core, 0); + eq_update_ci(eq->core, 0); set_ci = 0; } } - eq_update_ci(&eq->core, 1); + eq_update_ci(eq->core, 1); } static irqreturn_t mlx5_eq_pf_int(int irq, void *eq_ptr) @@ -368,6 +369,7 @@ static void eq_pf_action(struct work_struct *work) static int create_pf_eq(struct mlx5_core_dev *dev, struct mlx5_eq_pagefault *eq) { + struct mlx5_eq_param param = {}; int err; spin_lock_init(&eq->lock); @@ -386,11 +388,19 @@ create_pf_eq(struct mlx5_core_dev *dev, struct mlx5_eq_pagefault *eq) goto err_mempool; } - err = mlx5_create_async_eq(dev, &eq->core, MLX5_NUM_ASYNC_EQE, - 1 << MLX5_EVENT_TYPE_PAGE_FAULT, - "mlx5_page_fault_eq", mlx5_eq_pf_int); - if (err) + param = (struct mlx5_eq_param) { + .index = MLX5_EQ_PFAULT_IDX, + .mask = 1 << MLX5_EVENT_TYPE_PAGE_FAULT, + .nent = MLX5_NUM_ASYNC_EQE, + .context = eq, + .handler = mlx5_eq_pf_int + }; + + eq->core = mlx5_eq_create_generic(dev, "mlx5_page_fault_eq", ¶m); + if (IS_ERR(eq->core)) { + err = PTR_ERR(eq->core); goto err_wq; + } return 0; err_wq: @@ -404,7 +414,7 @@ static int destroy_pf_eq(struct mlx5_core_dev *dev, struct mlx5_eq_pagefault *eq { int err; - err = mlx5_destroy_async_eq(dev, &eq->core); + err = mlx5_eq_destroy_generic(dev, eq->core); cancel_work_sync(&eq->work); destroy_workqueue(eq->wq); mempool_destroy(eq->pool); @@ -710,25 +720,29 @@ static void init_eq_buf(struct mlx5_eq *eq) } static int -mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, - int nent, u64 mask, const char *name, irq_handler_t handler) +create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, const char *name, + struct mlx5_eq_param *param) { struct mlx5_eq_table *eq_table = dev->priv.eq_table; struct mlx5_cq_table *cq_table = &eq->cq_table; u32 out[MLX5_ST_SZ_DW(create_eq_out)] = {0}; struct mlx5_priv *priv = &dev->priv; + u8 vecidx = param->index; __be64 *pas; void *eqc; int inlen; u32 *in; int err; + if (eq_table->irq_info[vecidx].context) + return -EEXIST; + /* Init CQ table */ memset(cq_table, 0, sizeof(*cq_table)); spin_lock_init(&cq_table->lock); INIT_RADIX_TREE(&cq_table->tree, GFP_ATOMIC); - eq->nent = roundup_pow_of_two(nent + MLX5_NUM_SPARE_EQE); + eq->nent = roundup_pow_of_two(param->nent + MLX5_NUM_SPARE_EQE); eq->cons_index = 0; err = mlx5_buf_alloc(dev, eq->nent * MLX5_EQE_SIZE, &eq->buf); if (err) @@ -749,7 +763,7 @@ mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, mlx5_fill_page_array(&eq->buf, pas); MLX5_SET(create_eq_in, in, opcode, MLX5_CMD_OP_CREATE_EQ); - MLX5_SET64(create_eq_in, in, event_bitmask, mask); + MLX5_SET64(create_eq_in, in, event_bitmask, param->mask); eqc = MLX5_ADDR_OF(create_eq_in, in, eq_context_entry); MLX5_SET(eqc, eqc, log_eq_size, ilog2(eq->nent)); @@ -764,13 +778,15 @@ mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, snprintf(eq_table->irq_info[vecidx].name, MLX5_MAX_IRQ_NAME, "%s@pci:%s", name, pci_name(dev->pdev)); + eq_table->irq_info[vecidx].context = param->context; + eq->vecidx = vecidx; eq->eqn = MLX5_GET(create_eq_out, out, eq_number); eq->irqn = pci_irq_vector(dev->pdev, vecidx); eq->dev = dev; eq->doorbell = priv->uar->map + MLX5_EQ_DOORBEL_OFFSET; - err = request_irq(eq->irqn, handler, 0, - eq_table->irq_info[vecidx].name, eq); + err = request_irq(eq->irqn, param->handler, 0, + eq_table->irq_info[vecidx].name, param->context); if (err) goto err_eq; @@ -799,12 +815,19 @@ err_buf: return err; } -static int mlx5_destroy_unmap_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq) +static int destroy_unmap_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq) { + struct mlx5_eq_table *eq_table = dev->priv.eq_table; + struct mlx5_irq_info *irq_info; int err; + irq_info = &eq_table->irq_info[eq->vecidx]; + mlx5_debug_eq_remove(dev, eq); - free_irq(eq->irqn, eq); + + free_irq(eq->irqn, irq_info->context); + irq_info->context = NULL; + err = mlx5_cmd_destroy_eq(dev, eq->eqn); if (err) mlx5_core_warn(dev, "failed to destroy a previously created eq: eqn %d\n", @@ -883,48 +906,38 @@ void mlx5_eq_table_cleanup(struct mlx5_core_dev *dev) /* Async EQs */ -int mlx5_create_async_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, - int nent, u64 mask, const char *name, irq_handler_t handler) +static int create_async_eq(struct mlx5_core_dev *dev, const char *name, + struct mlx5_eq *eq, struct mlx5_eq_param *param) { struct mlx5_eq_table *eq_table = dev->priv.eq_table; - u8 vecdix; int err; mutex_lock(&eq_table->lock); - if (eq_table->num_async_eqs >= MLX5_EQ_MAX_ASYNC_EQS) { + if (param->index >= MLX5_EQ_MAX_ASYNC_EQS) { err = -ENOSPC; goto unlock; } - vecdix = eq_table->num_async_eqs + 1; - - err = mlx5_create_map_eq(dev, eq, vecdix, nent, mask, name, handler); - if (!err) - eq_table->num_async_eqs++; - + err = create_map_eq(dev, eq, name, param); unlock: mutex_unlock(&eq_table->lock); return err; } -int mlx5_destroy_async_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq) +static int destroy_async_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq) { struct mlx5_eq_table *eq_table = dev->priv.eq_table; int err; mutex_lock(&eq_table->lock); - err = mlx5_destroy_unmap_eq(dev, eq); - if (!err) - eq_table->num_async_eqs--; + err = destroy_unmap_eq(dev, eq); mutex_unlock(&eq_table->lock); return err; } -static int create_async_eqs(struct mlx5_core_dev *dev) +static u64 gather_async_events_mask(struct mlx5_core_dev *dev) { - struct mlx5_eq_table *table = dev->priv.eq_table; u64 async_event_mask = MLX5_ASYNC_EVENT_MASK; - int err; if (MLX5_VPORT_MANAGER(dev)) async_event_mask |= (1ull << MLX5_EVENT_TYPE_NIC_VPORT_CHANGE); @@ -953,9 +966,23 @@ static int create_async_eqs(struct mlx5_core_dev *dev) if (MLX5_CAP_MCAM_REG(dev, tracer_registers)) async_event_mask |= (1ull << MLX5_EVENT_TYPE_DEVICE_TRACER); - err = mlx5_create_async_eq(dev, &table->cmd_eq, MLX5_NUM_CMD_EQE, - 1ull << MLX5_EVENT_TYPE_CMD, "mlx5_cmd_eq", - mlx5_eq_async_int); + return async_event_mask; +} + +static int create_async_eqs(struct mlx5_core_dev *dev) +{ + struct mlx5_eq_table *table = dev->priv.eq_table; + struct mlx5_eq_param param = {}; + int err; + + param = (struct mlx5_eq_param) { + .index = MLX5_EQ_CMD_IDX, + .mask = 1ull << MLX5_EVENT_TYPE_CMD, + .nent = MLX5_NUM_CMD_EQE, + .context = &table->cmd_eq, + .handler = mlx5_eq_async_int, + }; + err = create_async_eq(dev, "mlx5_cmd_eq", &table->cmd_eq, ¶m); if (err) { mlx5_core_warn(dev, "failed to create cmd EQ %d\n", err); return err; @@ -963,15 +990,27 @@ static int create_async_eqs(struct mlx5_core_dev *dev) mlx5_cmd_use_events(dev); - err = mlx5_create_async_eq(dev, &table->async_eq, MLX5_NUM_ASYNC_EQE, - async_event_mask, "mlx5_async_eq", mlx5_eq_async_int); + param = (struct mlx5_eq_param) { + .index = MLX5_EQ_ASYNC_IDX, + .mask = gather_async_events_mask(dev), + .nent = MLX5_NUM_ASYNC_EQE, + .context = &table->async_eq, + .handler = mlx5_eq_async_int, + }; + err = create_async_eq(dev, "mlx5_async_eq", &table->async_eq, ¶m); if (err) { mlx5_core_warn(dev, "failed to create async EQ %d\n", err); goto err1; } - err = mlx5_create_async_eq(dev, &table->pages_eq, /* TODO: sriov max_vf + */ 1, - 1 << MLX5_EVENT_TYPE_PAGE_REQUEST, "mlx5_pages_eq", mlx5_eq_async_int); + param = (struct mlx5_eq_param) { + .index = MLX5_EQ_PAGEREQ_IDX, + .mask = 1 << MLX5_EVENT_TYPE_PAGE_REQUEST, + .nent = /* TODO: sriov max_vf + */ 1, + .context = &table->pages_eq, + .handler = mlx5_eq_async_int, + }; + err = create_async_eq(dev, "mlx5_pages_eq", &table->pages_eq, ¶m); if (err) { mlx5_core_warn(dev, "failed to create pages EQ %d\n", err); goto err2; @@ -989,17 +1028,17 @@ static int create_async_eqs(struct mlx5_core_dev *dev) return err; err3: - mlx5_destroy_async_eq(dev, &table->pages_eq); + destroy_async_eq(dev, &table->pages_eq); #else return err; #endif err2: - mlx5_destroy_async_eq(dev, &table->async_eq); + destroy_async_eq(dev, &table->async_eq); err1: mlx5_cmd_use_polling(dev); - mlx5_destroy_async_eq(dev, &table->cmd_eq); + destroy_async_eq(dev, &table->cmd_eq); return err; } @@ -1017,18 +1056,18 @@ static void destroy_async_eqs(struct mlx5_core_dev *dev) } #endif - err = mlx5_destroy_async_eq(dev, &table->pages_eq); + err = destroy_async_eq(dev, &table->pages_eq); if (err) mlx5_core_err(dev, "failed to destroy pages eq, err(%d)\n", err); - err = mlx5_destroy_async_eq(dev, &table->async_eq); + err = destroy_async_eq(dev, &table->async_eq); if (err) mlx5_core_err(dev, "failed to destroy async eq, err(%d)\n", err); mlx5_cmd_use_polling(dev); - err = mlx5_destroy_async_eq(dev, &table->cmd_eq); + err = destroy_async_eq(dev, &table->cmd_eq); if (err) mlx5_core_err(dev, "failed to destroy command eq, err(%d)\n", err); @@ -1049,6 +1088,77 @@ void mlx5_eq_synchronize_cmd_irq(struct mlx5_core_dev *dev) synchronize_irq(dev->priv.eq_table->cmd_eq.irqn); } +/* Generic EQ API for mlx5_core consumers + * Needed For RDMA ODP EQ for now + */ +struct mlx5_eq * +mlx5_eq_create_generic(struct mlx5_core_dev *dev, const char *name, + struct mlx5_eq_param *param) +{ + struct mlx5_eq *eq = kvzalloc(sizeof(*eq), GFP_KERNEL); + int err; + + if (!eq) + return ERR_PTR(-ENOMEM); + + err = create_async_eq(dev, name, eq, param); + if (err) { + kvfree(eq); + eq = ERR_PTR(err); + } + + return eq; +} +EXPORT_SYMBOL(mlx5_eq_create_generic); + +int mlx5_eq_destroy_generic(struct mlx5_core_dev *dev, struct mlx5_eq *eq) +{ + int err; + + if (IS_ERR(eq)) + return -EINVAL; + + err = destroy_async_eq(dev, eq); + if (err) + goto out; + + kvfree(eq); +out: + return err; +} +EXPORT_SYMBOL(mlx5_eq_destroy_generic); + +struct mlx5_eqe *mlx5_eq_get_eqe(struct mlx5_eq *eq, u32 cc) +{ + u32 ci = eq->cons_index + cc; + struct mlx5_eqe *eqe; + + eqe = get_eqe(eq, ci & (eq->nent - 1)); + eqe = ((eqe->owner & 1) ^ !!(ci & eq->nent)) ? NULL : eqe; + /* Make sure we read EQ entry contents after we've + * checked the ownership bit. + */ + if (eqe) + dma_rmb(); + + return eqe; +} +EXPORT_SYMBOL(mlx5_eq_get_eqe); + +void mlx5_eq_update_ci(struct mlx5_eq *eq, u32 cc, bool arm) +{ + __be32 __iomem *addr = eq->doorbell + (arm ? 0 : 2); + u32 val; + + eq->cons_index += cc; + val = (eq->cons_index & 0xffffff) | (eq->eqn << 24); + + __raw_writel((__force u32)cpu_to_be32(val), addr); + /* We still want ordering, just not swabbing, so add a barrier */ + mb(); +} +EXPORT_SYMBOL(mlx5_eq_update_ci); + /* Completion EQs */ static int set_comp_irq_affinity_hint(struct mlx5_core_dev *mdev, int i) @@ -1127,7 +1237,7 @@ static void destroy_comp_eqs(struct mlx5_core_dev *dev) #endif list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) { list_del(&eq->list); - if (mlx5_destroy_unmap_eq(dev, &eq->core)) + if (destroy_unmap_eq(dev, &eq->core)) mlx5_core_warn(dev, "failed to destroy comp EQ 0x%x\n", eq->core.eqn); tasklet_disable(&eq->tasklet_ctx.task); @@ -1155,6 +1265,7 @@ static int create_comp_eqs(struct mlx5_core_dev *dev) #endif for (i = 0; i < ncomp_vec; i++) { int vecidx = i + MLX5_EQ_VEC_COMP_BASE; + struct mlx5_eq_param param = {}; eq = kzalloc(sizeof(*eq), GFP_KERNEL); if (!eq) { @@ -1172,8 +1283,14 @@ static int create_comp_eqs(struct mlx5_core_dev *dev) irq_cpu_rmap_add(table->rmap, pci_irq_vector(dev->pdev, vecidx)); #endif snprintf(name, MLX5_MAX_IRQ_NAME, "mlx5_comp%d", i); - err = mlx5_create_map_eq(dev, &eq->core, vecidx, nent, 0, - name, mlx5_eq_comp_int); + param = (struct mlx5_eq_param) { + .index = vecidx, + .mask = 0, + .nent = nent, + .context = &eq->core, + .handler = mlx5_eq_comp_int + }; + err = create_map_eq(dev, &eq->core, name, ¶m); if (err) { kfree(eq); goto clean; @@ -1257,7 +1374,7 @@ struct mlx5_eq_comp *mlx5_eqn2comp_eq(struct mlx5_core_dev *dev, int eqn) void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev) { struct mlx5_eq_table *table = dev->priv.eq_table; - struct mlx5_eq_comp *eq; + int i, max_eqs; clear_comp_irqs_affinity_hints(dev); @@ -1267,16 +1384,16 @@ void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev) table->rmap = NULL; } #endif - list_for_each_entry(eq, &table->comp_eqs_list, list) - free_irq(eq->core.irqn, eq); - free_irq(table->pages_eq.irqn, &table->pages_eq); - free_irq(table->async_eq.irqn, &table->async_eq); - free_irq(table->cmd_eq.irqn, &table->cmd_eq); -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - if (MLX5_CAP_GEN(dev, pg)) - free_irq(table->pfault_eq.core.irqn, &table->pfault_eq.core); -#endif + mutex_lock(&table->lock); /* sync with create/destroy_async_eq */ + max_eqs = table->num_comp_vectors + MLX5_EQ_VEC_COMP_BASE; + for (i = max_eqs - 1; i >= 0; i--) { + if (!table->irq_info[i].context) + continue; + free_irq(pci_irq_vector(dev->pdev, i), table->irq_info[i].context); + table->irq_info[i].context = NULL; + } + mutex_unlock(&table->lock); pci_free_irq_vectors(dev->pdev); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h index 706d58383dbd..db32057ad054 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h @@ -7,11 +7,6 @@ #define MLX5_MAX_IRQ_NAME (32) -enum { - MLX5_EQ_MAX_ASYNC_EQS = 4, /* mlx5_core needs at least 3 */ - MLX5_EQ_VEC_COMP_BASE = MLX5_EQ_MAX_ASYNC_EQS, -}; - struct mlx5_eq_tasklet { struct list_head list; struct list_head process_list; @@ -31,6 +26,7 @@ struct mlx5_eq { u32 cons_index; struct mlx5_frag_buf buf; int size; + unsigned int vecidx; unsigned int irqn; u8 eqn; int nent; @@ -44,7 +40,7 @@ struct mlx5_eq_comp { }; struct mlx5_eq_pagefault { - struct mlx5_eq core; /* Must be first */ + struct mlx5_eq *core; struct work_struct work; spinlock_t lock; /* Pagefaults spinlock */ struct workqueue_struct *wq; @@ -55,10 +51,6 @@ int mlx5_eq_table_init(struct mlx5_core_dev *dev); void mlx5_eq_table_cleanup(struct mlx5_core_dev *dev); int mlx5_eq_table_create(struct mlx5_core_dev *dev); void mlx5_eq_table_destroy(struct mlx5_core_dev *dev); -int mlx5_create_async_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, - int nent, u64 mask, const char *name, - irq_handler_t handler); -int mlx5_destroy_async_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq); int mlx5_eq_add_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq); int mlx5_eq_del_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq); diff --git a/include/linux/mlx5/eq.h b/include/linux/mlx5/eq.h new file mode 100644 index 000000000000..c733673ba5f6 --- /dev/null +++ b/include/linux/mlx5/eq.h @@ -0,0 +1,39 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* Copyright (c) 2018 Mellanox Technologies. */ + +#ifndef MLX5_CORE_EQ_H +#define MLX5_CORE_EQ_H + +#include + +enum { + MLX5_EQ_PAGEREQ_IDX = 0, + MLX5_EQ_CMD_IDX = 1, + MLX5_EQ_ASYNC_IDX = 2, + /* reserved to be used by mlx5_core ulps (mlx5e/mlx5_ib) */ + MLX5_EQ_PFAULT_IDX = 3, + MLX5_EQ_MAX_ASYNC_EQS, + /* completion eqs vector indices start here */ + MLX5_EQ_VEC_COMP_BASE = MLX5_EQ_MAX_ASYNC_EQS, +}; + +struct mlx5_eq; + +struct mlx5_eq_param { + u8 index; + int nent; + u64 mask; + void *context; + irq_handler_t handler; +}; + +struct mlx5_eq * +mlx5_eq_create_generic(struct mlx5_core_dev *dev, const char *name, + struct mlx5_eq_param *param); +int +mlx5_eq_destroy_generic(struct mlx5_core_dev *dev, struct mlx5_eq *eq); + +struct mlx5_eqe *mlx5_eq_get_eqe(struct mlx5_eq *eq, u32 cc); +void mlx5_eq_update_ci(struct mlx5_eq *eq, u32 cc, bool arm); + +#endif /* MLX5_CORE_EQ_H */ -- cgit v1.2.3-55-g7522 From d5d284b829a6eb7127df24d1bd3896a698981e62 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 19 Nov 2018 10:52:41 -0800 Subject: {net,IB}/mlx5: Move Page fault EQ and ODP logic to RDMA Use the new generic EQ API to move all ODP RDMA data structures and logic form mlx5 core driver into mlx5_ib driver. Signed-off-by: Saeed Mahameed Reviewed-by: Leon Romanovsky Reviewed-by: Tariq Toukan Acked-by: Jason Gunthorpe Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/mlx5/main.c | 10 +- drivers/infiniband/hw/mlx5/mlx5_ib.h | 15 +- drivers/infiniband/hw/mlx5/odp.c | 281 ++++++++++++++++++++- drivers/net/ethernet/mellanox/mlx5/core/dev.c | 34 --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 252 ------------------ drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h | 8 - drivers/net/ethernet/mellanox/mlx5/core/main.c | 17 +- .../net/ethernet/mellanox/mlx5/core/mlx5_core.h | 2 - include/linux/mlx5/driver.h | 49 ---- include/linux/mlx5/eq.h | 21 ++ 10 files changed, 308 insertions(+), 381 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index 6fbc0cba1bac..fcf4a0328a90 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -6040,6 +6040,11 @@ static int mlx5_ib_stage_odp_init(struct mlx5_ib_dev *dev) return mlx5_ib_odp_init_one(dev); } +void mlx5_ib_stage_odp_cleanup(struct mlx5_ib_dev *dev) +{ + mlx5_ib_odp_cleanup_one(dev); +} + int mlx5_ib_stage_counters_init(struct mlx5_ib_dev *dev) { if (MLX5_CAP_GEN(dev->mdev, max_qp_cnt)) { @@ -6225,7 +6230,7 @@ static const struct mlx5_ib_profile pf_profile = { mlx5_ib_stage_dev_res_cleanup), STAGE_CREATE(MLX5_IB_STAGE_ODP, mlx5_ib_stage_odp_init, - NULL), + mlx5_ib_stage_odp_cleanup), STAGE_CREATE(MLX5_IB_STAGE_COUNTERS, mlx5_ib_stage_counters_init, mlx5_ib_stage_counters_cleanup), @@ -6395,9 +6400,6 @@ static struct mlx5_interface mlx5_ib_interface = { .add = mlx5_ib_add, .remove = mlx5_ib_remove, .event = mlx5_ib_event, -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - .pfault = mlx5_ib_pfault, -#endif .protocol = MLX5_INTERFACE_PROTOCOL_IB, }; diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h index b651a7a6fde9..27999fd32356 100644 --- a/drivers/infiniband/hw/mlx5/mlx5_ib.h +++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h @@ -880,6 +880,15 @@ struct mlx5_ib_lb_state { bool enabled; }; +struct mlx5_ib_pf_eq { + struct mlx5_ib_dev *dev; + struct mlx5_eq *core; + struct work_struct work; + spinlock_t lock; /* Pagefaults spinlock */ + struct workqueue_struct *wq; + mempool_t *pool; +}; + struct mlx5_ib_dev { struct ib_device ib_dev; const struct uverbs_object_tree_def *driver_trees[7]; @@ -902,6 +911,8 @@ struct mlx5_ib_dev { #ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING struct ib_odp_caps odp_caps; u64 odp_max_size; + struct mlx5_ib_pf_eq odp_pf_eq; + /* * Sleepable RCU that prevents destruction of MRs while they are still * being used by a page fault handler. @@ -1158,9 +1169,8 @@ struct ib_mr *mlx5_ib_reg_dm_mr(struct ib_pd *pd, struct ib_dm *dm, #ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING void mlx5_ib_internal_fill_odp_caps(struct mlx5_ib_dev *dev); -void mlx5_ib_pfault(struct mlx5_core_dev *mdev, void *context, - struct mlx5_pagefault *pfault); int mlx5_ib_odp_init_one(struct mlx5_ib_dev *ibdev); +void mlx5_ib_odp_cleanup_one(struct mlx5_ib_dev *ibdev); int __init mlx5_ib_odp_init(void); void mlx5_ib_odp_cleanup(void); void mlx5_ib_invalidate_range(struct ib_umem_odp *umem_odp, unsigned long start, @@ -1175,6 +1185,7 @@ static inline void mlx5_ib_internal_fill_odp_caps(struct mlx5_ib_dev *dev) } static inline int mlx5_ib_odp_init_one(struct mlx5_ib_dev *ibdev) { return 0; } +static inline void mlx5_ib_odp_cleanup_one(struct mlx5_ib_dev *ibdev) {} static inline int mlx5_ib_odp_init(void) { return 0; } static inline void mlx5_ib_odp_cleanup(void) {} static inline void mlx5_odp_init_mr_cache_entry(struct mlx5_cache_ent *ent) {} diff --git a/drivers/infiniband/hw/mlx5/odp.c b/drivers/infiniband/hw/mlx5/odp.c index 7d784b40e017..416d141322a0 100644 --- a/drivers/infiniband/hw/mlx5/odp.c +++ b/drivers/infiniband/hw/mlx5/odp.c @@ -37,6 +37,46 @@ #include "mlx5_ib.h" #include "cmd.h" +#include + +/* Contains the details of a pagefault. */ +struct mlx5_pagefault { + u32 bytes_committed; + u32 token; + u8 event_subtype; + u8 type; + union { + /* Initiator or send message responder pagefault details. */ + struct { + /* Received packet size, only valid for responders. */ + u32 packet_size; + /* + * Number of resource holding WQE, depends on type. + */ + u32 wq_num; + /* + * WQE index. Refers to either the send queue or + * receive queue, according to event_subtype. + */ + u16 wqe_index; + } wqe; + /* RDMA responder pagefault details */ + struct { + u32 r_key; + /* + * Received packet size, minimal size page fault + * resolution required for forward progress. + */ + u32 packet_size; + u32 rdma_op_len; + u64 rdma_va; + } rdma; + }; + + struct mlx5_ib_pf_eq *eq; + struct work_struct work; +}; + #define MAX_PREFETCH_LEN (4*1024*1024U) /* Timeout in ms to wait for an active mmu notifier to complete when handling @@ -304,14 +344,20 @@ static void mlx5_ib_page_fault_resume(struct mlx5_ib_dev *dev, { int wq_num = pfault->event_subtype == MLX5_PFAULT_SUBTYPE_WQE ? pfault->wqe.wq_num : pfault->token; - int ret = mlx5_core_page_fault_resume(dev->mdev, - pfault->token, - wq_num, - pfault->type, - error); - if (ret) - mlx5_ib_err(dev, "Failed to resolve the page fault on WQ 0x%x\n", - wq_num); + u32 out[MLX5_ST_SZ_DW(page_fault_resume_out)] = { }; + u32 in[MLX5_ST_SZ_DW(page_fault_resume_in)] = { }; + int err; + + MLX5_SET(page_fault_resume_in, in, opcode, MLX5_CMD_OP_PAGE_FAULT_RESUME); + MLX5_SET(page_fault_resume_in, in, page_fault_type, pfault->type); + MLX5_SET(page_fault_resume_in, in, token, pfault->token); + MLX5_SET(page_fault_resume_in, in, wq_number, wq_num); + MLX5_SET(page_fault_resume_in, in, error, !!error); + + err = mlx5_cmd_exec(dev->mdev, in, sizeof(in), out, sizeof(out)); + if (err) + mlx5_ib_err(dev, "Failed to resolve the page fault on WQ 0x%x err %d\n", + wq_num, err); } static struct mlx5_ib_mr *implicit_mr_alloc(struct ib_pd *pd, @@ -1196,10 +1242,8 @@ static void mlx5_ib_mr_rdma_pfault_handler(struct mlx5_ib_dev *dev, } } -void mlx5_ib_pfault(struct mlx5_core_dev *mdev, void *context, - struct mlx5_pagefault *pfault) +static void mlx5_ib_pfault(struct mlx5_ib_dev *dev, struct mlx5_pagefault *pfault) { - struct mlx5_ib_dev *dev = context; u8 event_subtype = pfault->event_subtype; switch (event_subtype) { @@ -1216,6 +1260,203 @@ void mlx5_ib_pfault(struct mlx5_core_dev *mdev, void *context, } } +static void mlx5_ib_eqe_pf_action(struct work_struct *work) +{ + struct mlx5_pagefault *pfault = container_of(work, + struct mlx5_pagefault, + work); + struct mlx5_ib_pf_eq *eq = pfault->eq; + + mlx5_ib_pfault(eq->dev, pfault); + mempool_free(pfault, eq->pool); +} + +static void mlx5_ib_eq_pf_process(struct mlx5_ib_pf_eq *eq) +{ + struct mlx5_eqe_page_fault *pf_eqe; + struct mlx5_pagefault *pfault; + struct mlx5_eqe *eqe; + int cc = 0; + + while ((eqe = mlx5_eq_get_eqe(eq->core, cc))) { + pfault = mempool_alloc(eq->pool, GFP_ATOMIC); + if (!pfault) { + schedule_work(&eq->work); + break; + } + + pf_eqe = &eqe->data.page_fault; + pfault->event_subtype = eqe->sub_type; + pfault->bytes_committed = be32_to_cpu(pf_eqe->bytes_committed); + + mlx5_ib_dbg(eq->dev, + "PAGE_FAULT: subtype: 0x%02x, bytes_committed: 0x%06x\n", + eqe->sub_type, pfault->bytes_committed); + + switch (eqe->sub_type) { + case MLX5_PFAULT_SUBTYPE_RDMA: + /* RDMA based event */ + pfault->type = + be32_to_cpu(pf_eqe->rdma.pftype_token) >> 24; + pfault->token = + be32_to_cpu(pf_eqe->rdma.pftype_token) & + MLX5_24BIT_MASK; + pfault->rdma.r_key = + be32_to_cpu(pf_eqe->rdma.r_key); + pfault->rdma.packet_size = + be16_to_cpu(pf_eqe->rdma.packet_length); + pfault->rdma.rdma_op_len = + be32_to_cpu(pf_eqe->rdma.rdma_op_len); + pfault->rdma.rdma_va = + be64_to_cpu(pf_eqe->rdma.rdma_va); + mlx5_ib_dbg(eq->dev, + "PAGE_FAULT: type:0x%x, token: 0x%06x, r_key: 0x%08x\n", + pfault->type, pfault->token, + pfault->rdma.r_key); + mlx5_ib_dbg(eq->dev, + "PAGE_FAULT: rdma_op_len: 0x%08x, rdma_va: 0x%016llx\n", + pfault->rdma.rdma_op_len, + pfault->rdma.rdma_va); + break; + + case MLX5_PFAULT_SUBTYPE_WQE: + /* WQE based event */ + pfault->type = + (be32_to_cpu(pf_eqe->wqe.pftype_wq) >> 24) & 0x7; + pfault->token = + be32_to_cpu(pf_eqe->wqe.token); + pfault->wqe.wq_num = + be32_to_cpu(pf_eqe->wqe.pftype_wq) & + MLX5_24BIT_MASK; + pfault->wqe.wqe_index = + be16_to_cpu(pf_eqe->wqe.wqe_index); + pfault->wqe.packet_size = + be16_to_cpu(pf_eqe->wqe.packet_length); + mlx5_ib_dbg(eq->dev, + "PAGE_FAULT: type:0x%x, token: 0x%06x, wq_num: 0x%06x, wqe_index: 0x%04x\n", + pfault->type, pfault->token, + pfault->wqe.wq_num, + pfault->wqe.wqe_index); + break; + + default: + mlx5_ib_warn(eq->dev, + "Unsupported page fault event sub-type: 0x%02hhx\n", + eqe->sub_type); + /* Unsupported page faults should still be + * resolved by the page fault handler + */ + } + + pfault->eq = eq; + INIT_WORK(&pfault->work, mlx5_ib_eqe_pf_action); + queue_work(eq->wq, &pfault->work); + + cc = mlx5_eq_update_cc(eq->core, ++cc); + } + + mlx5_eq_update_ci(eq->core, cc, 1); +} + +static irqreturn_t mlx5_ib_eq_pf_int(int irq, void *eq_ptr) +{ + struct mlx5_ib_pf_eq *eq = eq_ptr; + unsigned long flags; + + if (spin_trylock_irqsave(&eq->lock, flags)) { + mlx5_ib_eq_pf_process(eq); + spin_unlock_irqrestore(&eq->lock, flags); + } else { + schedule_work(&eq->work); + } + + return IRQ_HANDLED; +} + +/* mempool_refill() was proposed but unfortunately wasn't accepted + * http://lkml.iu.edu/hypermail/linux/kernel/1512.1/05073.html + * Cheap workaround. + */ +static void mempool_refill(mempool_t *pool) +{ + while (pool->curr_nr < pool->min_nr) + mempool_free(mempool_alloc(pool, GFP_KERNEL), pool); +} + +static void mlx5_ib_eq_pf_action(struct work_struct *work) +{ + struct mlx5_ib_pf_eq *eq = + container_of(work, struct mlx5_ib_pf_eq, work); + + mempool_refill(eq->pool); + + spin_lock_irq(&eq->lock); + mlx5_ib_eq_pf_process(eq); + spin_unlock_irq(&eq->lock); +} + +enum { + MLX5_IB_NUM_PF_EQE = 0x1000, + MLX5_IB_NUM_PF_DRAIN = 64, +}; + +static int +mlx5_ib_create_pf_eq(struct mlx5_ib_dev *dev, struct mlx5_ib_pf_eq *eq) +{ + struct mlx5_eq_param param = {}; + int err; + + INIT_WORK(&eq->work, mlx5_ib_eq_pf_action); + spin_lock_init(&eq->lock); + eq->dev = dev; + + eq->pool = mempool_create_kmalloc_pool(MLX5_IB_NUM_PF_DRAIN, + sizeof(struct mlx5_pagefault)); + if (!eq->pool) + return -ENOMEM; + + eq->wq = alloc_workqueue("mlx5_ib_page_fault", + WQ_HIGHPRI | WQ_UNBOUND | WQ_MEM_RECLAIM, + MLX5_NUM_CMD_EQE); + if (!eq->wq) { + err = -ENOMEM; + goto err_mempool; + } + + param = (struct mlx5_eq_param) { + .index = MLX5_EQ_PFAULT_IDX, + .mask = 1 << MLX5_EVENT_TYPE_PAGE_FAULT, + .nent = MLX5_IB_NUM_PF_EQE, + .context = eq, + .handler = mlx5_ib_eq_pf_int + }; + eq->core = mlx5_eq_create_generic(dev->mdev, "mlx5_ib_page_fault_eq", ¶m); + if (IS_ERR(eq->core)) { + err = PTR_ERR(eq->core); + goto err_wq; + } + + return 0; +err_wq: + destroy_workqueue(eq->wq); +err_mempool: + mempool_destroy(eq->pool); + return err; +} + +static int +mlx5_ib_destroy_pf_eq(struct mlx5_ib_dev *dev, struct mlx5_ib_pf_eq *eq) +{ + int err; + + err = mlx5_eq_destroy_generic(dev->mdev, eq->core); + cancel_work_sync(&eq->work); + destroy_workqueue(eq->wq); + mempool_destroy(eq->pool); + + return err; +} + void mlx5_odp_init_mr_cache_entry(struct mlx5_cache_ent *ent) { if (!(ent->dev->odp_caps.general_caps & IB_ODP_SUPPORT_IMPLICIT)) @@ -1244,7 +1485,7 @@ void mlx5_odp_init_mr_cache_entry(struct mlx5_cache_ent *ent) int mlx5_ib_odp_init_one(struct mlx5_ib_dev *dev) { - int ret; + int ret = 0; if (dev->odp_caps.general_caps & IB_ODP_SUPPORT_IMPLICIT) { ret = mlx5_cmd_null_mkey(dev->mdev, &dev->null_mkey); @@ -1254,7 +1495,20 @@ int mlx5_ib_odp_init_one(struct mlx5_ib_dev *dev) } } - return 0; + if (!MLX5_CAP_GEN(dev->mdev, pg)) + return ret; + + ret = mlx5_ib_create_pf_eq(dev, &dev->odp_pf_eq); + + return ret; +} + +void mlx5_ib_odp_cleanup_one(struct mlx5_ib_dev *dev) +{ + if (!MLX5_CAP_GEN(dev->mdev, pg)) + return; + + mlx5_ib_destroy_pf_eq(dev, &dev->odp_pf_eq); } int mlx5_ib_odp_init(void) @@ -1264,4 +1518,3 @@ int mlx5_ib_odp_init(void) return 0; } - diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c index 37ba7c78859d..7eedbea38a78 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c @@ -139,17 +139,6 @@ void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv) spin_lock_irq(&priv->ctx_lock); list_add_tail(&dev_ctx->list, &priv->ctx_list); - -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - if (dev_ctx->intf->pfault) { - if (priv->pfault) { - mlx5_core_err(dev, "multiple page fault handlers not supported"); - } else { - priv->pfault_ctx = dev_ctx->context; - priv->pfault = dev_ctx->intf->pfault; - } - } -#endif spin_unlock_irq(&priv->ctx_lock); } @@ -179,15 +168,6 @@ void mlx5_remove_device(struct mlx5_interface *intf, struct mlx5_priv *priv) if (!dev_ctx) return; -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - spin_lock_irq(&priv->ctx_lock); - if (priv->pfault == dev_ctx->intf->pfault) - priv->pfault = NULL; - spin_unlock_irq(&priv->ctx_lock); - - synchronize_srcu(&priv->pfault_srcu); -#endif - spin_lock_irq(&priv->ctx_lock); list_del(&dev_ctx->list); spin_unlock_irq(&priv->ctx_lock); @@ -447,20 +427,6 @@ void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event, spin_unlock_irqrestore(&priv->ctx_lock, flags); } -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING -void mlx5_core_page_fault(struct mlx5_core_dev *dev, - struct mlx5_pagefault *pfault) -{ - struct mlx5_priv *priv = &dev->priv; - int srcu_idx; - - srcu_idx = srcu_read_lock(&priv->pfault_srcu); - if (priv->pfault) - priv->pfault(dev, priv->pfault_ctx, pfault); - srcu_read_unlock(&priv->pfault_srcu, srcu_idx); -} -#endif - void mlx5_dev_list_lock(void) { mutex_lock(&mlx5_intf_mutex); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index ec1f5018546e..895401609c63 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -56,13 +56,6 @@ enum { MLX5_EQ_STATE_ALWAYS_ARMED = 0xb, }; -enum { - MLX5_NUM_SPARE_EQE = 0x80, - MLX5_NUM_ASYNC_EQE = 0x1000, - MLX5_NUM_CMD_EQE = 32, - MLX5_NUM_PF_DRAIN = 64, -}; - enum { MLX5_EQ_DOORBEL_OFFSET = 0x40, }; @@ -79,9 +72,6 @@ struct mlx5_eq_table { struct mlx5_eq async_eq; struct mlx5_eq cmd_eq; -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - struct mlx5_eq_pagefault pfault_eq; -#endif struct mutex lock; /* sync async eqs creations */ int num_comp_vectors; struct mlx5_irq_info *irq_info; @@ -222,224 +212,6 @@ static void eq_update_ci(struct mlx5_eq *eq, int arm) mb(); } -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING -static void eqe_pf_action(struct work_struct *work) -{ - struct mlx5_pagefault *pfault = container_of(work, - struct mlx5_pagefault, - work); - struct mlx5_eq_pagefault *eq = pfault->eq; - - mlx5_core_page_fault(eq->core->dev, pfault); - mempool_free(pfault, eq->pool); -} - -static void eq_pf_process(struct mlx5_eq_pagefault *eq) -{ - struct mlx5_core_dev *dev = eq->core->dev; - struct mlx5_eqe_page_fault *pf_eqe; - struct mlx5_pagefault *pfault; - struct mlx5_eqe *eqe; - int set_ci = 0; - - while ((eqe = next_eqe_sw(eq->core))) { - pfault = mempool_alloc(eq->pool, GFP_ATOMIC); - if (!pfault) { - schedule_work(&eq->work); - break; - } - - dma_rmb(); - pf_eqe = &eqe->data.page_fault; - pfault->event_subtype = eqe->sub_type; - pfault->bytes_committed = be32_to_cpu(pf_eqe->bytes_committed); - - mlx5_core_dbg(dev, - "PAGE_FAULT: subtype: 0x%02x, bytes_committed: 0x%06x\n", - eqe->sub_type, pfault->bytes_committed); - - switch (eqe->sub_type) { - case MLX5_PFAULT_SUBTYPE_RDMA: - /* RDMA based event */ - pfault->type = - be32_to_cpu(pf_eqe->rdma.pftype_token) >> 24; - pfault->token = - be32_to_cpu(pf_eqe->rdma.pftype_token) & - MLX5_24BIT_MASK; - pfault->rdma.r_key = - be32_to_cpu(pf_eqe->rdma.r_key); - pfault->rdma.packet_size = - be16_to_cpu(pf_eqe->rdma.packet_length); - pfault->rdma.rdma_op_len = - be32_to_cpu(pf_eqe->rdma.rdma_op_len); - pfault->rdma.rdma_va = - be64_to_cpu(pf_eqe->rdma.rdma_va); - mlx5_core_dbg(dev, - "PAGE_FAULT: type:0x%x, token: 0x%06x, r_key: 0x%08x\n", - pfault->type, pfault->token, - pfault->rdma.r_key); - mlx5_core_dbg(dev, - "PAGE_FAULT: rdma_op_len: 0x%08x, rdma_va: 0x%016llx\n", - pfault->rdma.rdma_op_len, - pfault->rdma.rdma_va); - break; - - case MLX5_PFAULT_SUBTYPE_WQE: - /* WQE based event */ - pfault->type = - (be32_to_cpu(pf_eqe->wqe.pftype_wq) >> 24) & 0x7; - pfault->token = - be32_to_cpu(pf_eqe->wqe.token); - pfault->wqe.wq_num = - be32_to_cpu(pf_eqe->wqe.pftype_wq) & - MLX5_24BIT_MASK; - pfault->wqe.wqe_index = - be16_to_cpu(pf_eqe->wqe.wqe_index); - pfault->wqe.packet_size = - be16_to_cpu(pf_eqe->wqe.packet_length); - mlx5_core_dbg(dev, - "PAGE_FAULT: type:0x%x, token: 0x%06x, wq_num: 0x%06x, wqe_index: 0x%04x\n", - pfault->type, pfault->token, - pfault->wqe.wq_num, - pfault->wqe.wqe_index); - break; - - default: - mlx5_core_warn(dev, - "Unsupported page fault event sub-type: 0x%02hhx\n", - eqe->sub_type); - /* Unsupported page faults should still be - * resolved by the page fault handler - */ - } - - pfault->eq = eq; - INIT_WORK(&pfault->work, eqe_pf_action); - queue_work(eq->wq, &pfault->work); - - ++eq->core->cons_index; - ++set_ci; - - if (unlikely(set_ci >= MLX5_NUM_SPARE_EQE)) { - eq_update_ci(eq->core, 0); - set_ci = 0; - } - } - - eq_update_ci(eq->core, 1); -} - -static irqreturn_t mlx5_eq_pf_int(int irq, void *eq_ptr) -{ - struct mlx5_eq_pagefault *eq = eq_ptr; - unsigned long flags; - - if (spin_trylock_irqsave(&eq->lock, flags)) { - eq_pf_process(eq); - spin_unlock_irqrestore(&eq->lock, flags); - } else { - schedule_work(&eq->work); - } - - return IRQ_HANDLED; -} - -/* mempool_refill() was proposed but unfortunately wasn't accepted - * http://lkml.iu.edu/hypermail/linux/kernel/1512.1/05073.html - * Chip workaround. - */ -static void mempool_refill(mempool_t *pool) -{ - while (pool->curr_nr < pool->min_nr) - mempool_free(mempool_alloc(pool, GFP_KERNEL), pool); -} - -static void eq_pf_action(struct work_struct *work) -{ - struct mlx5_eq_pagefault *eq = - container_of(work, struct mlx5_eq_pagefault, work); - - mempool_refill(eq->pool); - - spin_lock_irq(&eq->lock); - eq_pf_process(eq); - spin_unlock_irq(&eq->lock); -} - -static int -create_pf_eq(struct mlx5_core_dev *dev, struct mlx5_eq_pagefault *eq) -{ - struct mlx5_eq_param param = {}; - int err; - - spin_lock_init(&eq->lock); - INIT_WORK(&eq->work, eq_pf_action); - - eq->pool = mempool_create_kmalloc_pool(MLX5_NUM_PF_DRAIN, - sizeof(struct mlx5_pagefault)); - if (!eq->pool) - return -ENOMEM; - - eq->wq = alloc_workqueue("mlx5_page_fault", - WQ_HIGHPRI | WQ_UNBOUND | WQ_MEM_RECLAIM, - MLX5_NUM_CMD_EQE); - if (!eq->wq) { - err = -ENOMEM; - goto err_mempool; - } - - param = (struct mlx5_eq_param) { - .index = MLX5_EQ_PFAULT_IDX, - .mask = 1 << MLX5_EVENT_TYPE_PAGE_FAULT, - .nent = MLX5_NUM_ASYNC_EQE, - .context = eq, - .handler = mlx5_eq_pf_int - }; - - eq->core = mlx5_eq_create_generic(dev, "mlx5_page_fault_eq", ¶m); - if (IS_ERR(eq->core)) { - err = PTR_ERR(eq->core); - goto err_wq; - } - - return 0; -err_wq: - destroy_workqueue(eq->wq); -err_mempool: - mempool_destroy(eq->pool); - return err; -} - -static int destroy_pf_eq(struct mlx5_core_dev *dev, struct mlx5_eq_pagefault *eq) -{ - int err; - - err = mlx5_eq_destroy_generic(dev, eq->core); - cancel_work_sync(&eq->work); - destroy_workqueue(eq->wq); - mempool_destroy(eq->pool); - - return err; -} - -int mlx5_core_page_fault_resume(struct mlx5_core_dev *dev, u32 token, - u32 wq_num, u8 type, int error) -{ - u32 out[MLX5_ST_SZ_DW(page_fault_resume_out)] = {0}; - u32 in[MLX5_ST_SZ_DW(page_fault_resume_in)] = {0}; - - MLX5_SET(page_fault_resume_in, in, opcode, - MLX5_CMD_OP_PAGE_FAULT_RESUME); - MLX5_SET(page_fault_resume_in, in, error, !!error); - MLX5_SET(page_fault_resume_in, in, page_fault_type, type); - MLX5_SET(page_fault_resume_in, in, wq_number, wq_num); - MLX5_SET(page_fault_resume_in, in, token, token); - - return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); -} -EXPORT_SYMBOL_GPL(mlx5_core_page_fault_resume); -#endif - static void general_event_handler(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe) { @@ -1016,22 +788,7 @@ static int create_async_eqs(struct mlx5_core_dev *dev) goto err2; } -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - if (MLX5_CAP_GEN(dev, pg)) { - err = create_pf_eq(dev, &table->pfault_eq); - if (err) { - mlx5_core_warn(dev, "failed to create page fault EQ %d\n", - err); - goto err3; - } - } - - return err; -err3: - destroy_async_eq(dev, &table->pages_eq); -#else return err; -#endif err2: destroy_async_eq(dev, &table->async_eq); @@ -1047,15 +804,6 @@ static void destroy_async_eqs(struct mlx5_core_dev *dev) struct mlx5_eq_table *table = dev->priv.eq_table; int err; -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - if (MLX5_CAP_GEN(dev, pg)) { - err = destroy_pf_eq(dev, &table->pfault_eq); - if (err) - mlx5_core_err(dev, "failed to destroy page fault eq, err(%d)\n", - err); - } -#endif - err = destroy_async_eq(dev, &table->pages_eq); if (err) mlx5_core_err(dev, "failed to destroy pages eq, err(%d)\n", diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h index db32057ad054..4cc2d442cef6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h @@ -39,14 +39,6 @@ struct mlx5_eq_comp { struct list_head list; }; -struct mlx5_eq_pagefault { - struct mlx5_eq *core; - struct work_struct work; - spinlock_t lock; /* Pagefaults spinlock */ - struct workqueue_struct *wq; - mempool_t *pool; -}; - int mlx5_eq_table_init(struct mlx5_core_dev *dev); void mlx5_eq_table_cleanup(struct mlx5_core_dev *dev); int mlx5_eq_table_create(struct mlx5_core_dev *dev); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 3de83fe65f2b..91022f141855 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -1169,14 +1169,6 @@ static int init_one(struct pci_dev *pdev, INIT_LIST_HEAD(&priv->waiting_events_list); priv->is_accum_events = false; -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - err = init_srcu_struct(&priv->pfault_srcu); - if (err) { - dev_err(&pdev->dev, "init_srcu_struct failed with error code %d\n", - err); - goto clean_dev; - } -#endif mutex_init(&priv->bfregs.reg_head.lock); mutex_init(&priv->bfregs.wc_head.lock); INIT_LIST_HEAD(&priv->bfregs.reg_head.list); @@ -1185,7 +1177,7 @@ static int init_one(struct pci_dev *pdev, err = mlx5_pci_init(dev, priv); if (err) { dev_err(&pdev->dev, "mlx5_pci_init failed with error code %d\n", err); - goto clean_srcu; + goto clean_dev; } err = mlx5_health_init(dev); @@ -1218,11 +1210,7 @@ clean_health: mlx5_health_cleanup(dev); close_pci: mlx5_pci_close(dev, priv); -clean_srcu: -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - cleanup_srcu_struct(&priv->pfault_srcu); clean_dev: -#endif devlink_free(devlink); return err; @@ -1246,9 +1234,6 @@ static void remove_one(struct pci_dev *pdev) mlx5_pagealloc_cleanup(dev); mlx5_health_cleanup(dev); mlx5_pci_close(dev, priv); -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - cleanup_srcu_struct(&priv->pfault_srcu); -#endif devlink_free(devlink); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index 4728b027cb9e..21727d9eeb84 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -100,8 +100,6 @@ int mlx5_cmd_fast_teardown_hca(struct mlx5_core_dev *dev); void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event, unsigned long param); -void mlx5_core_page_fault(struct mlx5_core_dev *dev, - struct mlx5_pagefault *pfault); void mlx5_port_module_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe); void mlx5_enter_error_state(struct mlx5_core_dev *dev, bool force); void mlx5_disable_device(struct mlx5_core_dev *dev); diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index fe9b552aa649..f41e6713df10 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -510,7 +510,6 @@ struct mlx5_fc_stats { struct mlx5_mpfs; struct mlx5_eswitch; struct mlx5_lag; -struct mlx5_pagefault; struct mlx5_eq_table; struct mlx5_rate_limit { @@ -619,13 +618,6 @@ struct mlx5_priv { struct mlx5_port_module_event_stats pme_stats; -#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING - void (*pfault)(struct mlx5_core_dev *dev, - void *context, - struct mlx5_pagefault *pfault); - void *pfault_ctx; - struct srcu_struct pfault_srcu; -#endif struct mlx5_bfreg_data bfregs; struct mlx5_uars_page *uar; }; @@ -650,44 +642,6 @@ enum mlx5_pagefault_type_flags { MLX5_PFAULT_RDMA = 1 << 2, }; -/* Contains the details of a pagefault. */ -struct mlx5_pagefault { - u32 bytes_committed; - u32 token; - u8 event_subtype; - u8 type; - union { - /* Initiator or send message responder pagefault details. */ - struct { - /* Received packet size, only valid for responders. */ - u32 packet_size; - /* - * Number of resource holding WQE, depends on type. - */ - u32 wq_num; - /* - * WQE index. Refers to either the send queue or - * receive queue, according to event_subtype. - */ - u16 wqe_index; - } wqe; - /* RDMA responder pagefault details */ - struct { - u32 r_key; - /* - * Received packet size, minimal size page fault - * resolution required for forward progress. - */ - u32 packet_size; - u32 rdma_op_len; - u64 rdma_va; - } rdma; - }; - - struct mlx5_eq_pagefault *eq; - struct work_struct work; -}; - struct mlx5_td { struct list_head tirs_list; u32 tdn; @@ -1118,9 +1072,6 @@ struct mlx5_interface { void (*detach)(struct mlx5_core_dev *dev, void *context); void (*event)(struct mlx5_core_dev *dev, void *context, enum mlx5_dev_event event, unsigned long param); - void (*pfault)(struct mlx5_core_dev *dev, - void *context, - struct mlx5_pagefault *pfault); void * (*get_dev)(void *context); int protocol; struct list_head list; diff --git a/include/linux/mlx5/eq.h b/include/linux/mlx5/eq.h index c733673ba5f6..71d82c5a1a02 100644 --- a/include/linux/mlx5/eq.h +++ b/include/linux/mlx5/eq.h @@ -17,6 +17,10 @@ enum { MLX5_EQ_VEC_COMP_BASE = MLX5_EQ_MAX_ASYNC_EQS, }; +#define MLX5_NUM_CMD_EQE (32) +#define MLX5_NUM_ASYNC_EQE (0x1000) +#define MLX5_NUM_SPARE_EQE (0x80) + struct mlx5_eq; struct mlx5_eq_param { @@ -36,4 +40,21 @@ mlx5_eq_destroy_generic(struct mlx5_core_dev *dev, struct mlx5_eq *eq); struct mlx5_eqe *mlx5_eq_get_eqe(struct mlx5_eq *eq, u32 cc); void mlx5_eq_update_ci(struct mlx5_eq *eq, u32 cc, bool arm); +/* The HCA will think the queue has overflowed if we + * don't tell it we've been processing events. We + * create EQs with MLX5_NUM_SPARE_EQE extra entries, + * so we must update our consumer index at + * least that often. + * + * mlx5_eq_update_cc must be called on every EQE @EQ irq handler + */ +static inline u32 mlx5_eq_update_cc(struct mlx5_eq *eq, u32 cc) +{ + if (unlikely(cc >= MLX5_NUM_SPARE_EQE)) { + mlx5_eq_update_ci(eq, cc, 0); + cc = 0; + } + return cc; +} + #endif /* MLX5_CORE_EQ_H */ -- cgit v1.2.3-55-g7522 From 6d2d6fc83a281d51863fb5d08b59333ed1b46cc1 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 19 Nov 2018 10:52:42 -0800 Subject: net/mlx5: EQ, Make EQE access methods inline These are one/two liner generic EQ access methods, better have them declared static inline in eq.h. Signed-off-by: Saeed Mahameed Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 23 ---------------------- drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h | 25 +++++++++++++++++++++++- 2 files changed, 24 insertions(+), 24 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 895401609c63..6ba8e401a0c7 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -46,7 +46,6 @@ #include "diag/fw_tracer.h" enum { - MLX5_EQE_SIZE = sizeof(struct mlx5_eqe), MLX5_EQE_OWNER_INIT_VAL = 0x1, }; @@ -103,18 +102,6 @@ static int mlx5_cmd_destroy_eq(struct mlx5_core_dev *dev, u8 eqn) return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); } -static struct mlx5_eqe *get_eqe(struct mlx5_eq *eq, u32 entry) -{ - return mlx5_buf_offset(&eq->buf, entry * MLX5_EQE_SIZE); -} - -static struct mlx5_eqe *next_eqe_sw(struct mlx5_eq *eq) -{ - struct mlx5_eqe *eqe = get_eqe(eq, eq->cons_index & (eq->nent - 1)); - - return ((eqe->owner & 1) ^ !!(eq->cons_index & eq->nent)) ? NULL : eqe; -} - static const char *eqe_type_str(u8 type) { switch (type) { @@ -202,16 +189,6 @@ static enum mlx5_dev_event port_subtype_event(u8 subtype) return -1; } -static void eq_update_ci(struct mlx5_eq *eq, int arm) -{ - __be32 __iomem *addr = eq->doorbell + (arm ? 0 : 2); - u32 val = (eq->cons_index & 0xffffff) | (eq->eqn << 24); - - __raw_writel((__force u32)cpu_to_be32(val), addr); - /* We still want ordering, just not swabbing, so add a barrier */ - mb(); -} - static void general_event_handler(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe) { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h index 4cc2d442cef6..6d8c8a57d52b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h @@ -5,7 +5,8 @@ #define __LIB_MLX5_EQ_H__ #include -#define MLX5_MAX_IRQ_NAME (32) +#define MLX5_MAX_IRQ_NAME (32) +#define MLX5_EQE_SIZE (sizeof(struct mlx5_eqe)) struct mlx5_eq_tasklet { struct list_head list; @@ -39,6 +40,28 @@ struct mlx5_eq_comp { struct list_head list; }; +static inline struct mlx5_eqe *get_eqe(struct mlx5_eq *eq, u32 entry) +{ + return mlx5_buf_offset(&eq->buf, entry * MLX5_EQE_SIZE); +} + +static inline struct mlx5_eqe *next_eqe_sw(struct mlx5_eq *eq) +{ + struct mlx5_eqe *eqe = get_eqe(eq, eq->cons_index & (eq->nent - 1)); + + return ((eqe->owner & 1) ^ !!(eq->cons_index & eq->nent)) ? NULL : eqe; +} + +static inline void eq_update_ci(struct mlx5_eq *eq, int arm) +{ + __be32 __iomem *addr = eq->doorbell + (arm ? 0 : 2); + u32 val = (eq->cons_index & 0xffffff) | (eq->eqn << 24); + + __raw_writel((__force u32)cpu_to_be32(val), addr); + /* We still want ordering, just not swabbing, so add a barrier */ + mb(); +} + int mlx5_eq_table_init(struct mlx5_core_dev *dev); void mlx5_eq_table_cleanup(struct mlx5_core_dev *dev); int mlx5_eq_table_create(struct mlx5_core_dev *dev); -- cgit v1.2.3-55-g7522 From 6c0fbd726229fabecf8e8b631e3e134713eb30ef Mon Sep 17 00:00:00 2001 From: MichaÅ‚ MirosÅ‚aw Date: Tue, 20 Nov 2018 13:20:33 +0100 Subject: mlx5: use skb_vlan_tag_get_prio() Signed-off-by: MichaÅ‚ MirosÅ‚aw Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx5/core/en_tx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c index 6dacaeba2fbf..9afdf955f2bc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c @@ -127,7 +127,7 @@ u16 mlx5e_select_queue(struct net_device *dev, struct sk_buff *skb, else #endif if (skb_vlan_tag_present(skb)) - up = skb->vlan_tci >> VLAN_PRIO_SHIFT; + up = skb_vlan_tag_get_prio(skb); /* channel_ix can be larger than num_channels since * dev->num_real_tx_queues = num_channels * num_tc -- cgit v1.2.3-55-g7522 From 0f597ed435b9ea1296e25474b762bedceba97a50 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Tue, 20 Nov 2018 14:12:18 -0800 Subject: net/mlx5: EQ, Introduce atomic notifier chain subscription API Use atomic_notifier_chain to fire firmware events at internal mlx5 core components such as eswitch/fpga/clock/FW tracer/etc.., this is to avoid explicit calls from low level mlx5_core to upper components and to simplify the mlx5_core API for future developments. Simply provide register/unregister notifiers API and call the notifier chain on firmware async events. Example: to subscribe to a FW event: struct mlx5_nb port_event; MLX5_NB_INIT(&port_event, port_event_handler, PORT_CHANGE); mlx5_eq_notifier_register(mdev, &port_event); where: - port_event_handler is the notifier block callback. - PORT_EVENT is the suffix of MLX5_EVENT_TYPE_PORT_CHANGE. The above will guarantee that port_event_handler will receive all FW events of the type MLX5_EVENT_TYPE_PORT_CHANGE. To receive all FW/HW events one can subscribe to MLX5_EVENT_TYPE_NOTIFY_ANY. The next few patches will start moving all mlx5 core components to use this new API and cleanup mlx5_eq_async_int misx handler from component explicit calls and specific logic. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 42 ++++++++++++++++++++-- drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h | 5 +++ .../net/ethernet/mellanox/mlx5/core/mlx5_core.h | 5 +++ include/linux/mlx5/device.h | 10 +++++- include/linux/mlx5/eq.h | 16 +++++++-- 5 files changed, 72 insertions(+), 6 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 6ba8e401a0c7..34e4b2c246ff 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -31,6 +31,7 @@ */ #include +#include #include #include #include @@ -68,8 +69,10 @@ struct mlx5_irq_info { struct mlx5_eq_table { struct list_head comp_eqs_list; struct mlx5_eq pages_eq; - struct mlx5_eq async_eq; struct mlx5_eq cmd_eq; + struct mlx5_eq async_eq; + + struct atomic_notifier_head nh[MLX5_EVENT_TYPE_MAX]; struct mutex lock; /* sync async eqs creations */ int num_comp_vectors; @@ -316,13 +319,17 @@ u32 mlx5_eq_poll_irq_disabled(struct mlx5_eq_comp *eq) static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) { struct mlx5_eq *eq = eq_ptr; - struct mlx5_core_dev *dev = eq->dev; + struct mlx5_eq_table *eqt; + struct mlx5_core_dev *dev; struct mlx5_eqe *eqe; int set_ci = 0; u32 cqn = -1; u32 rsn; u8 port; + dev = eq->dev; + eqt = dev->priv.eq_table; + while ((eqe = next_eqe_sw(eq))) { /* * Make sure we read EQ entry contents after we've @@ -437,6 +444,13 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) break; } + if (likely(eqe->type < MLX5_EVENT_TYPE_MAX)) + atomic_notifier_call_chain(&eqt->nh[eqe->type], eqe->type, eqe); + else + mlx5_core_warn_once(dev, "notifier_call_chain is not setup for eqe: %d\n", eqe->type); + + atomic_notifier_call_chain(&eqt->nh[MLX5_EVENT_TYPE_NOTIFY_ANY], eqe->type, eqe); + ++eq->cons_index; ++set_ci; @@ -625,7 +639,7 @@ int mlx5_eq_del_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq) int mlx5_eq_table_init(struct mlx5_core_dev *dev) { struct mlx5_eq_table *eq_table; - int err; + int i, err; eq_table = kvzalloc(sizeof(*eq_table), GFP_KERNEL); if (!eq_table) @@ -638,6 +652,8 @@ int mlx5_eq_table_init(struct mlx5_core_dev *dev) goto kvfree_eq_table; mutex_init(&eq_table->lock); + for (i = 0; i < MLX5_EVENT_TYPE_MAX; i++) + ATOMIC_INIT_NOTIFIER_HEAD(&eq_table->nh[i]); return 0; @@ -1202,3 +1218,23 @@ void mlx5_eq_table_destroy(struct mlx5_core_dev *dev) destroy_async_eqs(dev); free_irq_vectors(dev); } + +int mlx5_eq_notifier_register(struct mlx5_core_dev *dev, struct mlx5_nb *nb) +{ + struct mlx5_eq_table *eqt = dev->priv.eq_table; + + if (nb->event_type >= MLX5_EVENT_TYPE_MAX) + return -EINVAL; + + return atomic_notifier_chain_register(&eqt->nh[nb->event_type], &nb->nb); +} + +int mlx5_eq_notifier_unregister(struct mlx5_core_dev *dev, struct mlx5_nb *nb) +{ + struct mlx5_eq_table *eqt = dev->priv.eq_table; + + if (nb->event_type >= MLX5_EVENT_TYPE_MAX) + return -EINVAL; + + return atomic_notifier_chain_unregister(&eqt->nh[nb->event_type], &nb->nb); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h index 6d8c8a57d52b..c0fb6d72b695 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/eq.h @@ -4,6 +4,8 @@ #ifndef __LIB_MLX5_EQ_H__ #define __LIB_MLX5_EQ_H__ #include +#include +#include #define MLX5_MAX_IRQ_NAME (32) #define MLX5_EQE_SIZE (sizeof(struct mlx5_eqe)) @@ -90,4 +92,7 @@ void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev); struct cpu_rmap *mlx5_eq_table_get_rmap(struct mlx5_core_dev *dev); #endif +int mlx5_eq_notifier_register(struct mlx5_core_dev *dev, struct mlx5_nb *nb); +int mlx5_eq_notifier_unregister(struct mlx5_core_dev *dev, struct mlx5_nb *nb); + #endif diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index 21727d9eeb84..e06c6e16ffc9 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -78,6 +78,11 @@ do { \ __func__, __LINE__, current->pid, \ ##__VA_ARGS__) +#define mlx5_core_warn_once(__dev, format, ...) \ + dev_warn_once(&(__dev)->pdev->dev, "%s:%d:(pid %d): " format, \ + __func__, __LINE__, current->pid, \ + ##__VA_ARGS__) + #define mlx5_core_info(__dev, format, ...) \ dev_info(&(__dev)->pdev->dev, format, ##__VA_ARGS__) diff --git a/include/linux/mlx5/device.h b/include/linux/mlx5/device.h index e326524bafcc..f7c8bebfe472 100644 --- a/include/linux/mlx5/device.h +++ b/include/linux/mlx5/device.h @@ -301,9 +301,15 @@ enum { MLX5_EVENT_QUEUE_TYPE_DCT = 6, }; +/* mlx5 components can subscribe to any one of these events via + * mlx5_eq_notifier_register API. + */ enum mlx5_event { + /* Special value to subscribe to any event */ + MLX5_EVENT_TYPE_NOTIFY_ANY = 0x0, + /* HW events enum start: comp events are not subscribable */ MLX5_EVENT_TYPE_COMP = 0x0, - + /* HW Async events enum start: subscribable events */ MLX5_EVENT_TYPE_PATH_MIG = 0x01, MLX5_EVENT_TYPE_COMM_EST = 0x02, MLX5_EVENT_TYPE_SQ_DRAINED = 0x03, @@ -341,6 +347,8 @@ enum mlx5_event { MLX5_EVENT_TYPE_FPGA_QP_ERROR = 0x21, MLX5_EVENT_TYPE_DEVICE_TRACER = 0x26, + + MLX5_EVENT_TYPE_MAX = MLX5_EVENT_TYPE_DEVICE_TRACER + 1, }; enum { diff --git a/include/linux/mlx5/eq.h b/include/linux/mlx5/eq.h index 71d82c5a1a02..00045cc4ea11 100644 --- a/include/linux/mlx5/eq.h +++ b/include/linux/mlx5/eq.h @@ -4,8 +4,6 @@ #ifndef MLX5_CORE_EQ_H #define MLX5_CORE_EQ_H -#include - enum { MLX5_EQ_PAGEREQ_IDX = 0, MLX5_EQ_CMD_IDX = 1, @@ -22,6 +20,7 @@ enum { #define MLX5_NUM_SPARE_EQE (0x80) struct mlx5_eq; +struct mlx5_core_dev; struct mlx5_eq_param { u8 index; @@ -57,4 +56,17 @@ static inline u32 mlx5_eq_update_cc(struct mlx5_eq *eq, u32 cc) return cc; } +struct mlx5_nb { + struct notifier_block nb; + u8 event_type; +}; + +#define mlx5_nb_cof(ptr, type, member) \ + (container_of(container_of(ptr, struct mlx5_nb, nb), type, member)) + +#define MLX5_NB_INIT(name, handler, event) do { \ + (name)->nb.notifier_call = handler; \ + (name)->event_type = MLX5_EVENT_TYPE_##event; \ +} while (0) + #endif /* MLX5_CORE_EQ_H */ -- cgit v1.2.3-55-g7522 From 720a936d407709dcbcf714293112667677d8c967 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Tue, 20 Nov 2018 14:12:19 -0800 Subject: net/mlx5: FWTrace, Use async events chain Remove the explicit call to mlx5_fw_tracer_event on MLX5_EVENT_TYPE_DEVICE_TRACER and let fw tracer to register its own handler when its ready. Signed-off-by: Saeed Mahameed --- .../ethernet/mellanox/mlx5/core/diag/fw_tracer.c | 27 ++++++++++++---------- .../ethernet/mellanox/mlx5/core/diag/fw_tracer.h | 2 +- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 4 ---- 3 files changed, 16 insertions(+), 17 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c b/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c index d4ec93bde4de..6999f4486e9e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c @@ -30,6 +30,7 @@ * SOFTWARE. */ #define CREATE_TRACE_POINTS +#include "lib/eq.h" #include "fw_tracer.h" #include "fw_tracer_tracepoint.h" @@ -846,9 +847,9 @@ free_tracer: return ERR_PTR(err); } -/* Create HW resources + start tracer - * must be called before Async EQ is created - */ +static int fw_tracer_event(struct notifier_block *nb, unsigned long action, void *data); + +/* Create HW resources + start tracer */ int mlx5_fw_tracer_init(struct mlx5_fw_tracer *tracer) { struct mlx5_core_dev *dev; @@ -874,6 +875,9 @@ int mlx5_fw_tracer_init(struct mlx5_fw_tracer *tracer) goto err_dealloc_pd; } + MLX5_NB_INIT(&tracer->nb, fw_tracer_event, DEVICE_TRACER); + mlx5_eq_notifier_register(dev, &tracer->nb); + mlx5_fw_tracer_start(tracer); return 0; @@ -883,9 +887,7 @@ err_dealloc_pd: return err; } -/* Stop tracer + Cleanup HW resources - * must be called after Async EQ is destroyed - */ +/* Stop tracer + Cleanup HW resources */ void mlx5_fw_tracer_cleanup(struct mlx5_fw_tracer *tracer) { if (IS_ERR_OR_NULL(tracer)) @@ -893,7 +895,7 @@ void mlx5_fw_tracer_cleanup(struct mlx5_fw_tracer *tracer) mlx5_core_dbg(tracer->dev, "FWTracer: Cleanup, is owner ? (%d)\n", tracer->owner); - + mlx5_eq_notifier_unregister(tracer->dev, &tracer->nb); cancel_work_sync(&tracer->ownership_change_work); cancel_work_sync(&tracer->handle_traces_work); @@ -922,12 +924,11 @@ void mlx5_fw_tracer_destroy(struct mlx5_fw_tracer *tracer) kfree(tracer); } -void mlx5_fw_tracer_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe) +static int fw_tracer_event(struct notifier_block *nb, unsigned long action, void *data) { - struct mlx5_fw_tracer *tracer = dev->tracer; - - if (!tracer) - return; + struct mlx5_fw_tracer *tracer = mlx5_nb_cof(nb, struct mlx5_fw_tracer, nb); + struct mlx5_core_dev *dev = tracer->dev; + struct mlx5_eqe *eqe = data; switch (eqe->sub_type) { case MLX5_TRACER_SUBTYPE_OWNERSHIP_CHANGE: @@ -942,6 +943,8 @@ void mlx5_fw_tracer_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe) mlx5_core_dbg(dev, "FWTracer: Event with unrecognized subtype: sub_type %d\n", eqe->sub_type); } + + return NOTIFY_OK; } EXPORT_TRACEPOINT_SYMBOL(mlx5_fw); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.h b/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.h index 0347f2dd5cee..a8b8747f2b61 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.h @@ -55,6 +55,7 @@ struct mlx5_fw_tracer { struct mlx5_core_dev *dev; + struct mlx5_nb nb; bool owner; u8 trc_ver; struct workqueue_struct *work_queue; @@ -170,6 +171,5 @@ struct mlx5_fw_tracer *mlx5_fw_tracer_create(struct mlx5_core_dev *dev); int mlx5_fw_tracer_init(struct mlx5_fw_tracer *tracer); void mlx5_fw_tracer_cleanup(struct mlx5_fw_tracer *tracer); void mlx5_fw_tracer_destroy(struct mlx5_fw_tracer *tracer); -void mlx5_fw_tracer_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe); #endif diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 34e4b2c246ff..c7c436b0ed2e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -434,10 +434,6 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) general_event_handler(dev, eqe); break; - case MLX5_EVENT_TYPE_DEVICE_TRACER: - mlx5_fw_tracer_event(dev, eqe); - break; - default: mlx5_core_warn(dev, "Unhandled event 0x%x on EQ 0x%x\n", eqe->type, eq->eqn); -- cgit v1.2.3-55-g7522 From a52a7d01fde117bc6d57602e2e7e947037c865b0 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Tue, 20 Nov 2018 14:12:20 -0800 Subject: net/mlx5: FPGA, Use async events chain Remove the explicit call to mlx5_fpga_event on MLX5_EVENT_TYPE_FPGA_ERROR or MLX5_EVENT_TYPE_FPGA_QP_ERROR let fpga core to register its own handler when its ready. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 5 --- .../net/ethernet/mellanox/mlx5/core/fpga/core.c | 38 +++++++++++++++++++--- .../net/ethernet/mellanox/mlx5/core/fpga/core.h | 11 +++---- 3 files changed, 38 insertions(+), 16 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index c7c436b0ed2e..8aabd23d2166 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -421,11 +421,6 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) mlx5_pps_event(dev, eqe); break; - case MLX5_EVENT_TYPE_FPGA_ERROR: - case MLX5_EVENT_TYPE_FPGA_QP_ERROR: - mlx5_fpga_event(dev, eqe->type, &eqe->data.raw); - break; - case MLX5_EVENT_TYPE_TEMP_WARN_EVENT: mlx5_temp_warning_event(dev, eqe); break; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.c index 436a8136f26f..27c5f6c7d36a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.c @@ -36,6 +36,7 @@ #include "mlx5_core.h" #include "lib/mlx5.h" +#include "lib/eq.h" #include "fpga/core.h" #include "fpga/conn.h" @@ -145,6 +146,22 @@ static int mlx5_fpga_device_brb(struct mlx5_fpga_device *fdev) return 0; } +static int mlx5_fpga_event(struct mlx5_fpga_device *, unsigned long, void *); + +static int fpga_err_event(struct notifier_block *nb, unsigned long event, void *eqe) +{ + struct mlx5_fpga_device *fdev = mlx5_nb_cof(nb, struct mlx5_fpga_device, fpga_err_nb); + + return mlx5_fpga_event(fdev, event, eqe); +} + +static int fpga_qp_err_event(struct notifier_block *nb, unsigned long event, void *eqe) +{ + struct mlx5_fpga_device *fdev = mlx5_nb_cof(nb, struct mlx5_fpga_device, fpga_qp_err_nb); + + return mlx5_fpga_event(fdev, event, eqe); +} + int mlx5_fpga_device_start(struct mlx5_core_dev *mdev) { struct mlx5_fpga_device *fdev = mdev->fpga; @@ -185,6 +202,11 @@ int mlx5_fpga_device_start(struct mlx5_core_dev *mdev) if (err) goto out; + MLX5_NB_INIT(&fdev->fpga_err_nb, fpga_err_event, FPGA_ERROR); + MLX5_NB_INIT(&fdev->fpga_qp_err_nb, fpga_qp_err_event, FPGA_QP_ERROR); + mlx5_eq_notifier_register(fdev->mdev, &fdev->fpga_err_nb); + mlx5_eq_notifier_register(fdev->mdev, &fdev->fpga_qp_err_nb); + err = mlx5_fpga_conn_device_init(fdev); if (err) goto err_rsvd_gid; @@ -201,6 +223,8 @@ err_conn_init: mlx5_fpga_conn_device_cleanup(fdev); err_rsvd_gid: + mlx5_eq_notifier_unregister(fdev->mdev, &fdev->fpga_err_nb); + mlx5_eq_notifier_unregister(fdev->mdev, &fdev->fpga_qp_err_nb); mlx5_core_unreserve_gids(mdev, max_num_qps); out: spin_lock_irqsave(&fdev->state_lock, flags); @@ -256,6 +280,9 @@ void mlx5_fpga_device_stop(struct mlx5_core_dev *mdev) } mlx5_fpga_conn_device_cleanup(fdev); + mlx5_eq_notifier_unregister(fdev->mdev, &fdev->fpga_err_nb); + mlx5_eq_notifier_unregister(fdev->mdev, &fdev->fpga_qp_err_nb); + max_num_qps = MLX5_CAP_FPGA(mdev, shell_caps.max_num_qps); mlx5_core_unreserve_gids(mdev, max_num_qps); } @@ -283,9 +310,10 @@ static const char *mlx5_fpga_qp_syndrome_to_string(u8 syndrome) return "Unknown"; } -void mlx5_fpga_event(struct mlx5_core_dev *mdev, u8 event, void *data) +static int mlx5_fpga_event(struct mlx5_fpga_device *fdev, + unsigned long event, void *eqe) { - struct mlx5_fpga_device *fdev = mdev->fpga; + void *data = ((struct mlx5_eqe *)eqe)->data.raw; const char *event_name; bool teardown = false; unsigned long flags; @@ -303,9 +331,7 @@ void mlx5_fpga_event(struct mlx5_core_dev *mdev, u8 event, void *data) fpga_qpn = MLX5_GET(fpga_qp_error_event, data, fpga_qpn); break; default: - mlx5_fpga_warn_ratelimited(fdev, "Unexpected event %u\n", - event); - return; + return NOTIFY_DONE; } spin_lock_irqsave(&fdev->state_lock, flags); @@ -326,4 +352,6 @@ void mlx5_fpga_event(struct mlx5_core_dev *mdev, u8 event, void *data) */ if (teardown) mlx5_trigger_health_work(fdev->mdev); + + return NOTIFY_OK; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.h index 3e2355c8df3f..7e2e871dbf83 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.h @@ -35,11 +35,16 @@ #ifdef CONFIG_MLX5_FPGA +#include + +#include "lib/eq.h" #include "fpga/cmd.h" /* Represents an Innova device */ struct mlx5_fpga_device { struct mlx5_core_dev *mdev; + struct mlx5_nb fpga_err_nb; + struct mlx5_nb fpga_qp_err_nb; spinlock_t state_lock; /* Protects state transitions */ enum mlx5_fpga_status state; enum mlx5_fpga_image last_admin_image; @@ -82,7 +87,6 @@ int mlx5_fpga_init(struct mlx5_core_dev *mdev); void mlx5_fpga_cleanup(struct mlx5_core_dev *mdev); int mlx5_fpga_device_start(struct mlx5_core_dev *mdev); void mlx5_fpga_device_stop(struct mlx5_core_dev *mdev); -void mlx5_fpga_event(struct mlx5_core_dev *mdev, u8 event, void *data); #else @@ -104,11 +108,6 @@ static inline void mlx5_fpga_device_stop(struct mlx5_core_dev *mdev) { } -static inline void mlx5_fpga_event(struct mlx5_core_dev *mdev, u8 event, - void *data) -{ -} - #endif #endif /* __MLX5_FPGA_CORE_H__ */ -- cgit v1.2.3-55-g7522 From 41069256e93045a45a2c359c9715439be0b47bf4 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Tue, 20 Nov 2018 14:12:21 -0800 Subject: net/mlx5: Clock, Use async events chain Remove the explicit call to mlx5_pps_event on MLX5_EVENT_TYPE_PPS_EVENT and let clock logic to register its own handler when its ready. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 4 ---- .../net/ethernet/mellanox/mlx5/core/lib/clock.c | 24 +++++++++++++++------- .../net/ethernet/mellanox/mlx5/core/lib/clock.h | 3 --- include/linux/mlx5/driver.h | 4 +++- 4 files changed, 20 insertions(+), 15 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 8aabd23d2166..e5fcce9ca107 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -417,10 +417,6 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) mlx5_port_module_event(dev, eqe); break; - case MLX5_EVENT_TYPE_PPS_EVENT: - mlx5_pps_event(dev, eqe); - break; - case MLX5_EVENT_TYPE_TEMP_WARN_EVENT: mlx5_temp_warning_event(dev, eqe); break; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c index 0d90b1b4a3d3..d27c239e7d6c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c @@ -33,6 +33,7 @@ #include #include #include +#include "lib/eq.h" #include "en.h" #include "clock.h" @@ -439,16 +440,17 @@ static void mlx5_get_pps_caps(struct mlx5_core_dev *mdev) clock->pps_info.pin_caps[7] = MLX5_GET(mtpps_reg, out, cap_pin_7_mode); } -void mlx5_pps_event(struct mlx5_core_dev *mdev, - struct mlx5_eqe *eqe) +static int mlx5_pps_event(struct notifier_block *nb, + unsigned long type, void *data) { - struct mlx5_clock *clock = &mdev->clock; + struct mlx5_clock *clock = mlx5_nb_cof(nb, struct mlx5_clock, pps_nb); + struct mlx5_core_dev *mdev = clock->mdev; struct ptp_clock_event ptp_event; - struct timespec64 ts; - u64 nsec_now, nsec_delta; u64 cycles_now, cycles_delta; + u64 nsec_now, nsec_delta, ns; + struct mlx5_eqe *eqe = data; int pin = eqe->data.pps.pin; - s64 ns; + struct timespec64 ts; unsigned long flags; switch (clock->ptp_info.pin_config[pin].func) { @@ -463,6 +465,7 @@ void mlx5_pps_event(struct mlx5_core_dev *mdev, } else { ptp_event.type = PTP_CLOCK_EXTTS; } + /* TODOL clock->ptp can be NULL if ptp_clock_register failes */ ptp_clock_event(clock->ptp, &ptp_event); break; case PTP_PF_PEROUT: @@ -481,8 +484,11 @@ void mlx5_pps_event(struct mlx5_core_dev *mdev, write_sequnlock_irqrestore(&clock->lock, flags); break; default: - mlx5_core_err(mdev, " Unhandled event\n"); + mlx5_core_err(mdev, " Unhandled clock PPS event, func %d\n", + clock->ptp_info.pin_config[pin].func); } + + return NOTIFY_OK; } void mlx5_init_clock(struct mlx5_core_dev *mdev) @@ -567,6 +573,9 @@ void mlx5_init_clock(struct mlx5_core_dev *mdev) PTR_ERR(clock->ptp)); clock->ptp = NULL; } + + MLX5_NB_INIT(&clock->pps_nb, mlx5_pps_event, PPS_EVENT); + mlx5_eq_notifier_register(mdev, &clock->pps_nb); } void mlx5_cleanup_clock(struct mlx5_core_dev *mdev) @@ -576,6 +585,7 @@ void mlx5_cleanup_clock(struct mlx5_core_dev *mdev) if (!MLX5_CAP_GEN(mdev, device_frequency_khz)) return; + mlx5_eq_notifier_unregister(mdev, &clock->pps_nb); if (clock->ptp) { ptp_clock_unregister(clock->ptp); clock->ptp = NULL; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.h index 263cb6e2aeee..31600924bdc3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.h @@ -36,7 +36,6 @@ #if IS_ENABLED(CONFIG_PTP_1588_CLOCK) void mlx5_init_clock(struct mlx5_core_dev *mdev); void mlx5_cleanup_clock(struct mlx5_core_dev *mdev); -void mlx5_pps_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe); static inline int mlx5_clock_get_ptp_index(struct mlx5_core_dev *mdev) { @@ -60,8 +59,6 @@ static inline ktime_t mlx5_timecounter_cyc2time(struct mlx5_clock *clock, #else static inline void mlx5_init_clock(struct mlx5_core_dev *mdev) {} static inline void mlx5_cleanup_clock(struct mlx5_core_dev *mdev) {} -static inline void mlx5_pps_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe) {} - static inline int mlx5_clock_get_ptp_index(struct mlx5_core_dev *mdev) { return -1; diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index f41e6713df10..99a23db9a929 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -50,6 +50,7 @@ #include #include #include +#include #include #include @@ -671,6 +672,8 @@ struct mlx5_pps { }; struct mlx5_clock { + struct mlx5_core_dev *mdev; + struct mlx5_nb pps_nb; seqlock_t lock; struct cyclecounter cycles; struct timecounter tc; @@ -678,7 +681,6 @@ struct mlx5_clock { u32 nominal_c_mult; unsigned long overflow_period; struct delayed_work overflow_work; - struct mlx5_core_dev *mdev; struct ptp_clock *ptp; struct ptp_clock_info ptp_info; struct mlx5_pps pps_info; -- cgit v1.2.3-55-g7522 From 6933a93795590b51a3056b6d66562ca1da1557ae Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Tue, 20 Nov 2018 14:12:22 -0800 Subject: net/mlx5: E-Switch, Use async events chain Remove the explicit call to mlx5_eswitch_vport_event on MLX5_EVENT_TYPE_NIC_VPORT_CHANGE and let the eswitch register its own handler when its ready. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 4 --- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 44 +++++++++++++---------- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 3 +- 3 files changed, 26 insertions(+), 25 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index e5fcce9ca107..7c8b2d89645b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -409,10 +409,6 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) } break; - case MLX5_EVENT_TYPE_NIC_VPORT_CHANGE: - mlx5_eswitch_vport_event(dev->priv.eswitch, eqe); - break; - case MLX5_EVENT_TYPE_PORT_MODULE_EVENT: mlx5_port_module_event(dev, eqe); break; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index 2346b6ba3d54..e6a9b19d8626 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -36,6 +36,7 @@ #include #include #include "mlx5_core.h" +#include "lib/eq.h" #include "eswitch.h" #include "fs_core.h" #include "lib/eq.h" @@ -1568,7 +1569,6 @@ static void esw_disable_vport(struct mlx5_eswitch *esw, int vport_num) /* Mark this vport as disabled to discard new events */ vport->enabled = false; - mlx5_eq_synchronize_async_irq(esw->dev); /* Wait for current already scheduled events to complete */ flush_workqueue(esw->work_queue); /* Disable events from this vport */ @@ -1594,10 +1594,25 @@ static void esw_disable_vport(struct mlx5_eswitch *esw, int vport_num) mutex_unlock(&esw->state_lock); } +static int eswitch_vport_event(struct notifier_block *nb, + unsigned long type, void *data) +{ + struct mlx5_eswitch *esw = mlx5_nb_cof(nb, struct mlx5_eswitch, nb); + struct mlx5_eqe *eqe = data; + struct mlx5_vport *vport; + u16 vport_num; + + vport_num = be16_to_cpu(eqe->data.vport_change.vport_num); + vport = &esw->vports[vport_num]; + if (vport->enabled) + queue_work(esw->work_queue, &vport->vport_change_handler); + + return NOTIFY_OK; +} + /* Public E-Switch API */ #define ESW_ALLOWED(esw) ((esw) && MLX5_ESWITCH_MANAGER((esw)->dev)) - int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode) { int err; @@ -1641,6 +1656,11 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode) for (i = 0; i <= nvfs; i++) esw_enable_vport(esw, i, enabled_events); + if (mode == SRIOV_LEGACY) { + MLX5_NB_INIT(&esw->nb, eswitch_vport_event, NIC_VPORT_CHANGE); + mlx5_eq_notifier_register(esw->dev, &esw->nb); + } + esw_info(esw->dev, "SRIOV enabled: active vports(%d)\n", esw->enabled_vports); return 0; @@ -1670,6 +1690,9 @@ void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw) mc_promisc = &esw->mc_promisc; nvports = esw->enabled_vports; + if (esw->mode == SRIOV_LEGACY) + mlx5_eq_notifier_unregister(esw->dev, &esw->nb); + for (i = 0; i < esw->total_vports; i++) esw_disable_vport(esw, i); @@ -1778,23 +1801,6 @@ void mlx5_eswitch_cleanup(struct mlx5_eswitch *esw) kfree(esw); } -void mlx5_eswitch_vport_event(struct mlx5_eswitch *esw, struct mlx5_eqe *eqe) -{ - struct mlx5_eqe_vport_change *vc_eqe = &eqe->data.vport_change; - u16 vport_num = be16_to_cpu(vc_eqe->vport_num); - struct mlx5_vport *vport; - - if (!esw) { - pr_warn("MLX5 E-Switch: vport %d got an event while eswitch is not initialized\n", - vport_num); - return; - } - - vport = &esw->vports[vport_num]; - if (vport->enabled) - queue_work(esw->work_queue, &vport->vport_change_handler); -} - /* Vport Administration */ #define LEGAL_VPORT(esw, vport) (vport >= 0 && vport < esw->total_vports) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h index aaafc9f17115..480ffa294867 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h @@ -181,6 +181,7 @@ struct esw_mc_addr { /* SRIOV only */ struct mlx5_eswitch { struct mlx5_core_dev *dev; + struct mlx5_nb nb; struct mlx5_eswitch_fdb fdb_table; struct hlist_head mc_table[MLX5_L2_ADDR_HASH_SIZE]; struct workqueue_struct *work_queue; @@ -211,7 +212,6 @@ int esw_offloads_init_reps(struct mlx5_eswitch *esw); /* E-Switch API */ int mlx5_eswitch_init(struct mlx5_core_dev *dev); void mlx5_eswitch_cleanup(struct mlx5_eswitch *esw); -void mlx5_eswitch_vport_event(struct mlx5_eswitch *esw, struct mlx5_eqe *eqe); int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode); void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw); int mlx5_eswitch_set_vport_mac(struct mlx5_eswitch *esw, @@ -352,7 +352,6 @@ static inline bool mlx5_eswitch_vlan_actions_supported(struct mlx5_core_dev *dev /* eswitch API stubs */ static inline int mlx5_eswitch_init(struct mlx5_core_dev *dev) { return 0; } static inline void mlx5_eswitch_cleanup(struct mlx5_eswitch *esw) {} -static inline void mlx5_eswitch_vport_event(struct mlx5_eswitch *esw, struct mlx5_eqe *eqe) {} static inline int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode) { return 0; } static inline void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw) {} -- cgit v1.2.3-55-g7522 From 0cf53c1247565b339a23d82a1853a0c41e9a2a34 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Tue, 20 Nov 2018 14:12:23 -0800 Subject: net/mlx5: FWPage, Use async events chain Remove the explicit call to mlx5_core_req_pages_handler on MLX5_EVENT_TYPE_PAGE_REQUEST and let FW page logic to register its own handler when its ready. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 11 ------ drivers/net/ethernet/mellanox/mlx5/core/main.c | 27 +++++++------ .../net/ethernet/mellanox/mlx5/core/pagealloc.c | 44 +++++++++++++++------- include/linux/mlx5/driver.h | 5 ++- 4 files changed, 47 insertions(+), 40 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 7c8b2d89645b..7f6a644700eb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -398,17 +398,6 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) mlx5_eq_cq_event(eq, cqn, eqe->type); break; - case MLX5_EVENT_TYPE_PAGE_REQUEST: - { - u16 func_id = be16_to_cpu(eqe->data.req_pages.func_id); - s32 npages = be32_to_cpu(eqe->data.req_pages.num_pages); - - mlx5_core_dbg(dev, "page request for func 0x%x, npages %d\n", - func_id, npages); - mlx5_core_req_pages_handler(dev, func_id, npages); - } - break; - case MLX5_EVENT_TYPE_PORT_MODULE_EVENT: mlx5_port_module_event(dev, eqe); break; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 91022f141855..9e4cd2757ea8 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -916,16 +916,10 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, goto reclaim_boot_pages; } - err = mlx5_pagealloc_start(dev); - if (err) { - dev_err(&pdev->dev, "mlx5_pagealloc_start failed\n"); - goto reclaim_boot_pages; - } - err = mlx5_cmd_init_hca(dev, sw_owner_id); if (err) { dev_err(&pdev->dev, "init hca failed\n"); - goto err_pagealloc_stop; + goto reclaim_boot_pages; } mlx5_set_driver_version(dev); @@ -953,6 +947,8 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, goto err_get_uars; } + mlx5_pagealloc_start(dev); + err = mlx5_eq_table_create(dev); if (err) { dev_err(&pdev->dev, "Failed to create EQs\n"); @@ -1039,6 +1035,7 @@ err_fw_tracer: mlx5_eq_table_destroy(dev); err_eq_table: + mlx5_pagealloc_stop(dev); mlx5_put_uars_page(dev, priv->uar); err_get_uars: @@ -1052,9 +1049,6 @@ err_stop_poll: goto out_err; } -err_pagealloc_stop: - mlx5_pagealloc_stop(dev); - reclaim_boot_pages: mlx5_reclaim_startup_pages(dev); @@ -1100,16 +1094,18 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, mlx5_fpga_device_stop(dev); mlx5_fw_tracer_cleanup(dev->tracer); mlx5_eq_table_destroy(dev); + mlx5_pagealloc_stop(dev); mlx5_put_uars_page(dev, priv->uar); + if (cleanup) mlx5_cleanup_once(dev); mlx5_stop_health_poll(dev, cleanup); + err = mlx5_cmd_teardown_hca(dev); if (err) { dev_err(&dev->pdev->dev, "tear_down_hca failed, skip cleanup\n"); goto out; } - mlx5_pagealloc_stop(dev); mlx5_reclaim_startup_pages(dev); mlx5_core_disable_hca(dev, 0); mlx5_cmd_cleanup(dev); @@ -1186,12 +1182,14 @@ static int init_one(struct pci_dev *pdev, goto close_pci; } - mlx5_pagealloc_init(dev); + err = mlx5_pagealloc_init(dev); + if (err) + goto err_pagealloc_init; err = mlx5_load_one(dev, priv, true); if (err) { dev_err(&pdev->dev, "mlx5_load_one failed with error code %d\n", err); - goto clean_health; + goto err_load_one; } request_module_nowait(MLX5_IB_MOD); @@ -1205,8 +1203,9 @@ static int init_one(struct pci_dev *pdev, clean_load: mlx5_unload_one(dev, priv, true); -clean_health: +err_load_one: mlx5_pagealloc_cleanup(dev); +err_pagealloc_init: mlx5_health_cleanup(dev); close_pci: mlx5_pci_close(dev, priv); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c index e36d3e3675f9..a83b517b0714 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c @@ -37,6 +37,7 @@ #include #include #include "mlx5_core.h" +#include "lib/eq.h" enum { MLX5_PAGES_CANT_GIVE = 0, @@ -433,15 +434,28 @@ static void pages_work_handler(struct work_struct *work) kfree(req); } -void mlx5_core_req_pages_handler(struct mlx5_core_dev *dev, u16 func_id, - s32 npages) +static int req_pages_handler(struct notifier_block *nb, + unsigned long type, void *data) { struct mlx5_pages_req *req; - + struct mlx5_core_dev *dev; + struct mlx5_priv *priv; + struct mlx5_eqe *eqe; + u16 func_id; + s32 npages; + + priv = mlx5_nb_cof(nb, struct mlx5_priv, pg_nb); + dev = container_of(priv, struct mlx5_core_dev, priv); + eqe = data; + + func_id = be16_to_cpu(eqe->data.req_pages.func_id); + npages = be32_to_cpu(eqe->data.req_pages.num_pages); + mlx5_core_dbg(dev, "page request for func 0x%x, npages %d\n", + func_id, npages); req = kzalloc(sizeof(*req), GFP_ATOMIC); if (!req) { mlx5_core_warn(dev, "failed to allocate pages request\n"); - return; + return NOTIFY_DONE; } req->dev = dev; @@ -449,6 +463,7 @@ void mlx5_core_req_pages_handler(struct mlx5_core_dev *dev, u16 func_id, req->npages = npages; INIT_WORK(&req->work, pages_work_handler); queue_work(dev->priv.pg_wq, &req->work); + return NOTIFY_OK; } int mlx5_satisfy_startup_pages(struct mlx5_core_dev *dev, int boot) @@ -524,29 +539,32 @@ int mlx5_reclaim_startup_pages(struct mlx5_core_dev *dev) return 0; } -void mlx5_pagealloc_init(struct mlx5_core_dev *dev) +int mlx5_pagealloc_init(struct mlx5_core_dev *dev) { dev->priv.page_root = RB_ROOT; INIT_LIST_HEAD(&dev->priv.free_list); + dev->priv.pg_wq = create_singlethread_workqueue("mlx5_page_allocator"); + if (!dev->priv.pg_wq) + return -ENOMEM; + + return 0; } void mlx5_pagealloc_cleanup(struct mlx5_core_dev *dev) { - /* nothing */ + destroy_workqueue(dev->priv.pg_wq); } -int mlx5_pagealloc_start(struct mlx5_core_dev *dev) +void mlx5_pagealloc_start(struct mlx5_core_dev *dev) { - dev->priv.pg_wq = create_singlethread_workqueue("mlx5_page_allocator"); - if (!dev->priv.pg_wq) - return -ENOMEM; - - return 0; + MLX5_NB_INIT(&dev->priv.pg_nb, req_pages_handler, PAGE_REQUEST); + mlx5_eq_notifier_register(dev, &dev->priv.pg_nb); } void mlx5_pagealloc_stop(struct mlx5_core_dev *dev) { - destroy_workqueue(dev->priv.pg_wq); + mlx5_eq_notifier_unregister(dev, &dev->priv.pg_nb); + flush_workqueue(dev->priv.pg_wq); } int mlx5_wait_for_vf_pages(struct mlx5_core_dev *dev) diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 99a23db9a929..61088ad33500 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -564,6 +564,7 @@ struct mlx5_priv { struct mlx5_eq_table *eq_table; /* pages stuff */ + struct mlx5_nb pg_nb; struct workqueue_struct *pg_wq; struct rb_root page_root; int fw_pages; @@ -962,9 +963,9 @@ int mlx5_core_alloc_pd(struct mlx5_core_dev *dev, u32 *pdn); int mlx5_core_dealloc_pd(struct mlx5_core_dev *dev, u32 pdn); int mlx5_core_mad_ifc(struct mlx5_core_dev *dev, const void *inb, void *outb, u16 opmod, u8 port); -void mlx5_pagealloc_init(struct mlx5_core_dev *dev); +int mlx5_pagealloc_init(struct mlx5_core_dev *dev); void mlx5_pagealloc_cleanup(struct mlx5_core_dev *dev); -int mlx5_pagealloc_start(struct mlx5_core_dev *dev); +void mlx5_pagealloc_start(struct mlx5_core_dev *dev); void mlx5_pagealloc_stop(struct mlx5_core_dev *dev); void mlx5_core_req_pages_handler(struct mlx5_core_dev *dev, u16 func_id, s32 npages); -- cgit v1.2.3-55-g7522 From 71edc69ca1a78ce18411a540c550a4ef1eb017cd Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Tue, 20 Nov 2018 14:12:24 -0800 Subject: net/mlx5: CmdIF, Use async events chain Remove the explicit call to mlx5_cmd_comp_handler on MLX5_EVENT_TYPE_CMD and let command interface to register its own handler when its ready. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/cmd.c | 48 +++++++++++++++++++++- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 4 -- drivers/net/ethernet/mellanox/mlx5/core/health.c | 25 +---------- .../net/ethernet/mellanox/mlx5/core/mlx5_core.h | 2 +- include/linux/mlx5/driver.h | 2 + 5 files changed, 50 insertions(+), 31 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c index 7b18aff955f1..8ab636d59edb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c @@ -40,9 +40,11 @@ #include #include #include +#include #include #include "mlx5_core.h" +#include "lib/eq.h" enum { CMD_IF_REV = 5, @@ -805,6 +807,8 @@ static u16 msg_to_opcode(struct mlx5_cmd_msg *in) return MLX5_GET(mbox_in, in->first.data, opcode); } +static void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool forced); + static void cb_timeout_handler(struct work_struct *work) { struct delayed_work *dwork = container_of(work, struct delayed_work, @@ -1412,14 +1416,32 @@ static void mlx5_cmd_change_mod(struct mlx5_core_dev *dev, int mode) up(&cmd->sem); } +static int cmd_comp_notifier(struct notifier_block *nb, + unsigned long type, void *data) +{ + struct mlx5_core_dev *dev; + struct mlx5_cmd *cmd; + struct mlx5_eqe *eqe; + + cmd = mlx5_nb_cof(nb, struct mlx5_cmd, nb); + dev = container_of(cmd, struct mlx5_core_dev, cmd); + eqe = data; + + mlx5_cmd_comp_handler(dev, be32_to_cpu(eqe->data.cmd.vector), false); + + return NOTIFY_OK; +} void mlx5_cmd_use_events(struct mlx5_core_dev *dev) { + MLX5_NB_INIT(&dev->cmd.nb, cmd_comp_notifier, CMD); + mlx5_eq_notifier_register(dev, &dev->cmd.nb); mlx5_cmd_change_mod(dev, CMD_MODE_EVENTS); } void mlx5_cmd_use_polling(struct mlx5_core_dev *dev) { mlx5_cmd_change_mod(dev, CMD_MODE_POLLING); + mlx5_eq_notifier_unregister(dev, &dev->cmd.nb); } static void free_msg(struct mlx5_core_dev *dev, struct mlx5_cmd_msg *msg) @@ -1435,7 +1457,7 @@ static void free_msg(struct mlx5_core_dev *dev, struct mlx5_cmd_msg *msg) } } -void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool forced) +static void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool forced) { struct mlx5_cmd *cmd = &dev->cmd; struct mlx5_cmd_work_ent *ent; @@ -1533,7 +1555,29 @@ void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool forced) } } } -EXPORT_SYMBOL(mlx5_cmd_comp_handler); + +void mlx5_cmd_trigger_completions(struct mlx5_core_dev *dev) +{ + unsigned long flags; + u64 vector; + + /* wait for pending handlers to complete */ + mlx5_eq_synchronize_cmd_irq(dev); + spin_lock_irqsave(&dev->cmd.alloc_lock, flags); + vector = ~dev->cmd.bitmask & ((1ul << (1 << dev->cmd.log_sz)) - 1); + if (!vector) + goto no_trig; + + vector |= MLX5_TRIGGERED_CMD_COMP; + spin_unlock_irqrestore(&dev->cmd.alloc_lock, flags); + + mlx5_core_dbg(dev, "vector 0x%llx\n", vector); + mlx5_cmd_comp_handler(dev, vector, true); + return; + +no_trig: + spin_unlock_irqrestore(&dev->cmd.alloc_lock, flags); +} static int status_to_err(u8 status) { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 7f6a644700eb..b28869aa1a4e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -368,10 +368,6 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) mlx5_srq_event(dev, rsn, eqe->type); break; - case MLX5_EVENT_TYPE_CMD: - mlx5_cmd_comp_handler(dev, be32_to_cpu(eqe->data.cmd.vector), false); - break; - case MLX5_EVENT_TYPE_PORT_CHANGE: port = (eqe->data.port.port >> 4) & 0xf; switch (eqe->sub_type) { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c index 066883003aea..4e42bd290959 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/health.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c @@ -79,29 +79,6 @@ void mlx5_set_nic_state(struct mlx5_core_dev *dev, u8 state) &dev->iseg->cmdq_addr_l_sz); } -static void trigger_cmd_completions(struct mlx5_core_dev *dev) -{ - unsigned long flags; - u64 vector; - - /* wait for pending handlers to complete */ - mlx5_eq_synchronize_cmd_irq(dev); - spin_lock_irqsave(&dev->cmd.alloc_lock, flags); - vector = ~dev->cmd.bitmask & ((1ul << (1 << dev->cmd.log_sz)) - 1); - if (!vector) - goto no_trig; - - vector |= MLX5_TRIGGERED_CMD_COMP; - spin_unlock_irqrestore(&dev->cmd.alloc_lock, flags); - - mlx5_core_dbg(dev, "vector 0x%llx\n", vector); - mlx5_cmd_comp_handler(dev, vector, true); - return; - -no_trig: - spin_unlock_irqrestore(&dev->cmd.alloc_lock, flags); -} - static int in_fatal(struct mlx5_core_dev *dev) { struct mlx5_core_health *health = &dev->priv.health; @@ -125,7 +102,7 @@ void mlx5_enter_error_state(struct mlx5_core_dev *dev, bool force) mlx5_core_err(dev, "start\n"); if (pci_channel_offline(dev->pdev) || in_fatal(dev) || force) { dev->state = MLX5_DEVICE_STATE_INTERNAL_ERROR; - trigger_cmd_completions(dev); + mlx5_cmd_trigger_completions(dev); } mlx5_core_event(dev, MLX5_DEV_EVENT_SYS_ERROR, 1); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index e06c6e16ffc9..5dd453e47a04 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -127,7 +127,7 @@ int mlx5_destroy_scheduling_element_cmd(struct mlx5_core_dev *dev, u8 hierarchy, int mlx5_wait_for_vf_pages(struct mlx5_core_dev *dev); u64 mlx5_read_internal_timer(struct mlx5_core_dev *dev); -void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool forced); +void mlx5_cmd_trigger_completions(struct mlx5_core_dev *dev); int mlx5_cq_debugfs_init(struct mlx5_core_dev *dev); void mlx5_cq_debugfs_cleanup(struct mlx5_core_dev *dev); diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 61088ad33500..a8d638134fc8 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -278,6 +278,8 @@ struct mlx5_cmd_stats { }; struct mlx5_cmd { + struct mlx5_nb nb; + void *cmd_alloc_buf; dma_addr_t alloc_dma; int alloc_size; -- cgit v1.2.3-55-g7522 From 221c14f3d12489ced0f2ca8b31b2221c5dbbf145 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Tue, 20 Nov 2018 14:12:25 -0800 Subject: net/mlx5: Resource tables, Use async events chain Remove the explicit call to QP/SRQ resources events handlers on several FW events and let resources logic register resources events notifiers via the new API. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 29 ------------ drivers/net/ethernet/mellanox/mlx5/core/qp.c | 68 ++++++++++++++++++++++----- drivers/net/ethernet/mellanox/mlx5/core/srq.c | 55 +++++++++++++++++++--- include/linux/mlx5/driver.h | 6 ++- 4 files changed, 108 insertions(+), 50 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index b28869aa1a4e..0cf448575ebd 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -324,7 +324,6 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) struct mlx5_eqe *eqe; int set_ci = 0; u32 cqn = -1; - u32 rsn; u8 port; dev = eq->dev; @@ -340,34 +339,6 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) mlx5_core_dbg(eq->dev, "eqn %d, eqe type %s\n", eq->eqn, eqe_type_str(eqe->type)); switch (eqe->type) { - case MLX5_EVENT_TYPE_DCT_DRAINED: - rsn = be32_to_cpu(eqe->data.dct.dctn) & 0xffffff; - rsn |= (MLX5_RES_DCT << MLX5_USER_INDEX_LEN); - mlx5_rsc_event(dev, rsn, eqe->type); - break; - case MLX5_EVENT_TYPE_PATH_MIG: - case MLX5_EVENT_TYPE_COMM_EST: - case MLX5_EVENT_TYPE_SQ_DRAINED: - case MLX5_EVENT_TYPE_SRQ_LAST_WQE: - case MLX5_EVENT_TYPE_WQ_CATAS_ERROR: - case MLX5_EVENT_TYPE_PATH_MIG_FAILED: - case MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR: - case MLX5_EVENT_TYPE_WQ_ACCESS_ERROR: - rsn = be32_to_cpu(eqe->data.qp_srq.qp_srq_n) & 0xffffff; - rsn |= (eqe->data.qp_srq.type << MLX5_USER_INDEX_LEN); - mlx5_core_dbg(dev, "event %s(%d) arrived on resource 0x%x\n", - eqe_type_str(eqe->type), eqe->type, rsn); - mlx5_rsc_event(dev, rsn, eqe->type); - break; - - case MLX5_EVENT_TYPE_SRQ_RQ_LIMIT: - case MLX5_EVENT_TYPE_SRQ_CATAS_ERROR: - rsn = be32_to_cpu(eqe->data.qp_srq.qp_srq_n) & 0xffffff; - mlx5_core_dbg(dev, "SRQ event %s(%d): srqn 0x%x\n", - eqe_type_str(eqe->type), eqe->type, rsn); - mlx5_srq_event(dev, rsn, eqe->type); - break; - case MLX5_EVENT_TYPE_PORT_CHANGE: port = (eqe->data.port.port >> 4) & 0xf; switch (eqe->sub_type) { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/qp.c b/drivers/net/ethernet/mellanox/mlx5/core/qp.c index cba4a435043a..28726c63101f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/qp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/qp.c @@ -38,11 +38,11 @@ #include #include "mlx5_core.h" +#include "lib/eq.h" -static struct mlx5_core_rsc_common *mlx5_get_rsc(struct mlx5_core_dev *dev, - u32 rsn) +static struct mlx5_core_rsc_common * +mlx5_get_rsc(struct mlx5_qp_table *table, u32 rsn) { - struct mlx5_qp_table *table = &dev->priv.qp_table; struct mlx5_core_rsc_common *common; spin_lock(&table->lock); @@ -53,11 +53,6 @@ static struct mlx5_core_rsc_common *mlx5_get_rsc(struct mlx5_core_dev *dev, spin_unlock(&table->lock); - if (!common) { - mlx5_core_warn(dev, "Async event for bogus resource 0x%x\n", - rsn); - return NULL; - } return common; } @@ -120,14 +115,52 @@ static bool is_event_type_allowed(int rsc_type, int event_type) } } -void mlx5_rsc_event(struct mlx5_core_dev *dev, u32 rsn, int event_type) +static int rsc_event_notifier(struct notifier_block *nb, + unsigned long type, void *data) { - struct mlx5_core_rsc_common *common = mlx5_get_rsc(dev, rsn); + struct mlx5_core_rsc_common *common; + struct mlx5_qp_table *table; + struct mlx5_core_dev *dev; struct mlx5_core_dct *dct; + u8 event_type = (u8)type; struct mlx5_core_qp *qp; + struct mlx5_priv *priv; + struct mlx5_eqe *eqe; + u32 rsn; + + switch (event_type) { + case MLX5_EVENT_TYPE_DCT_DRAINED: + eqe = data; + rsn = be32_to_cpu(eqe->data.dct.dctn) & 0xffffff; + rsn |= (MLX5_RES_DCT << MLX5_USER_INDEX_LEN); + break; + case MLX5_EVENT_TYPE_PATH_MIG: + case MLX5_EVENT_TYPE_COMM_EST: + case MLX5_EVENT_TYPE_SQ_DRAINED: + case MLX5_EVENT_TYPE_SRQ_LAST_WQE: + case MLX5_EVENT_TYPE_WQ_CATAS_ERROR: + case MLX5_EVENT_TYPE_PATH_MIG_FAILED: + case MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR: + case MLX5_EVENT_TYPE_WQ_ACCESS_ERROR: + eqe = data; + rsn = be32_to_cpu(eqe->data.qp_srq.qp_srq_n) & 0xffffff; + rsn |= (eqe->data.qp_srq.type << MLX5_USER_INDEX_LEN); + break; + default: + return NOTIFY_DONE; + } + + table = mlx5_nb_cof(nb, struct mlx5_qp_table, nb); + priv = container_of(table, struct mlx5_priv, qp_table); + dev = container_of(priv, struct mlx5_core_dev, priv); - if (!common) - return; + mlx5_core_dbg(dev, "event (%d) arrived on resource 0x%x\n", eqe->type, rsn); + + common = mlx5_get_rsc(table, rsn); + if (!common) { + mlx5_core_warn(dev, "Async event for bogus resource 0x%x\n", rsn); + return NOTIFY_OK; + } if (!is_event_type_allowed((rsn >> MLX5_USER_INDEX_LEN), event_type)) { mlx5_core_warn(dev, "event 0x%.2x is not allowed on resource 0x%.8x\n", @@ -152,6 +185,8 @@ void mlx5_rsc_event(struct mlx5_core_dev *dev, u32 rsn, int event_type) } out: mlx5_core_put_rsc(common); + + return NOTIFY_OK; } static int create_resource_common(struct mlx5_core_dev *dev, @@ -487,10 +522,16 @@ void mlx5_init_qp_table(struct mlx5_core_dev *dev) spin_lock_init(&table->lock); INIT_RADIX_TREE(&table->tree, GFP_ATOMIC); mlx5_qp_debugfs_init(dev); + + MLX5_NB_INIT(&table->nb, rsc_event_notifier, NOTIFY_ANY); + mlx5_eq_notifier_register(dev, &table->nb); } void mlx5_cleanup_qp_table(struct mlx5_core_dev *dev) { + struct mlx5_qp_table *table = &dev->priv.qp_table; + + mlx5_eq_notifier_unregister(dev, &table->nb); mlx5_qp_debugfs_cleanup(dev); } @@ -676,8 +717,9 @@ struct mlx5_core_rsc_common *mlx5_core_res_hold(struct mlx5_core_dev *dev, enum mlx5_res_type res_type) { u32 rsn = res_num | (res_type << MLX5_USER_INDEX_LEN); + struct mlx5_qp_table *table = &dev->priv.qp_table; - return mlx5_get_rsc(dev, rsn); + return mlx5_get_rsc(table, rsn); } EXPORT_SYMBOL_GPL(mlx5_core_res_hold); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/srq.c b/drivers/net/ethernet/mellanox/mlx5/core/srq.c index 6a6fc9be01e6..0563866c13f2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/srq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/srq.c @@ -36,13 +36,25 @@ #include #include #include -#include "mlx5_core.h" #include +#include "mlx5_core.h" +#include "lib/eq.h" -void mlx5_srq_event(struct mlx5_core_dev *dev, u32 srqn, int event_type) +static int srq_event_notifier(struct mlx5_srq_table *table, + unsigned long type, void *data) { - struct mlx5_srq_table *table = &dev->priv.srq_table; + struct mlx5_core_dev *dev; struct mlx5_core_srq *srq; + struct mlx5_priv *priv; + struct mlx5_eqe *eqe; + u32 srqn; + + priv = container_of(table, struct mlx5_priv, srq_table); + dev = container_of(priv, struct mlx5_core_dev, priv); + + eqe = data; + srqn = be32_to_cpu(eqe->data.qp_srq.qp_srq_n) & 0xffffff; + mlx5_core_dbg(dev, "SRQ event (%d): srqn 0x%x\n", eqe->type, srqn); spin_lock(&table->lock); @@ -54,13 +66,35 @@ void mlx5_srq_event(struct mlx5_core_dev *dev, u32 srqn, int event_type) if (!srq) { mlx5_core_warn(dev, "Async event for bogus SRQ 0x%08x\n", srqn); - return; + return NOTIFY_OK; } - srq->event(srq, event_type); + srq->event(srq, eqe->type); if (atomic_dec_and_test(&srq->refcount)) complete(&srq->free); + + return NOTIFY_OK; +} + +static int catas_err_notifier(struct notifier_block *nb, + unsigned long type, void *data) +{ + struct mlx5_srq_table *table; + + table = mlx5_nb_cof(nb, struct mlx5_srq_table, catas_err_nb); + /* type == MLX5_EVENT_TYPE_SRQ_CATAS_ERROR */ + return srq_event_notifier(table, type, data); +} + +static int rq_limit_notifier(struct notifier_block *nb, + unsigned long type, void *data) +{ + struct mlx5_srq_table *table; + + table = mlx5_nb_cof(nb, struct mlx5_srq_table, rq_limit_nb); + /* type == MLX5_EVENT_TYPE_SRQ_RQ_LIMIT */ + return srq_event_notifier(table, type, data); } static int get_pas_size(struct mlx5_srq_attr *in) @@ -708,9 +742,18 @@ void mlx5_init_srq_table(struct mlx5_core_dev *dev) memset(table, 0, sizeof(*table)); spin_lock_init(&table->lock); INIT_RADIX_TREE(&table->tree, GFP_ATOMIC); + + MLX5_NB_INIT(&table->catas_err_nb, catas_err_notifier, SRQ_CATAS_ERROR); + mlx5_eq_notifier_register(dev, &table->catas_err_nb); + + MLX5_NB_INIT(&table->rq_limit_nb, rq_limit_notifier, SRQ_RQ_LIMIT); + mlx5_eq_notifier_register(dev, &table->rq_limit_nb); } void mlx5_cleanup_srq_table(struct mlx5_core_dev *dev) { - /* nothing */ + struct mlx5_srq_table *table = &dev->priv.srq_table; + + mlx5_eq_notifier_unregister(dev, &table->rq_limit_nb); + mlx5_eq_notifier_unregister(dev, &table->catas_err_nb); } diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index a8d638134fc8..afba0864f45c 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -464,6 +464,8 @@ struct mlx5_core_health { }; struct mlx5_qp_table { + struct mlx5_nb nb; + /* protect radix tree */ spinlock_t lock; @@ -471,6 +473,8 @@ struct mlx5_qp_table { }; struct mlx5_srq_table { + struct mlx5_nb catas_err_nb; + struct mlx5_nb rq_limit_nb; /* protect radix tree */ spinlock_t lock; @@ -978,8 +982,6 @@ void mlx5_unregister_debugfs(void); void mlx5_fill_page_array(struct mlx5_frag_buf *buf, __be64 *pas); void mlx5_fill_page_frag_array(struct mlx5_frag_buf *frag_buf, __be64 *pas); -void mlx5_rsc_event(struct mlx5_core_dev *dev, u32 rsn, int event_type); -void mlx5_srq_event(struct mlx5_core_dev *dev, u32 srqn, int event_type); struct mlx5_core_srq *mlx5_core_get_srq(struct mlx5_core_dev *dev, u32 srqn); int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn, unsigned int *irqn); -- cgit v1.2.3-55-g7522 From 2742bc90bc102644d67e17758e234646d91ae696 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Tue, 20 Nov 2018 14:12:26 -0800 Subject: net/mlx5: CQ ERR, Use async events chain Remove the explicit call to mlx5_eq_cq_event on MLX5_EVENT_TYPE_CQ_ERROR and register a specific CQ ERROR handler via the new API. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 66 ++++++++++++++++++---------- 1 file changed, 44 insertions(+), 22 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 0cf448575ebd..4e3febbf639d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -74,6 +74,9 @@ struct mlx5_eq_table { struct atomic_notifier_head nh[MLX5_EVENT_TYPE_MAX]; + /* Since CQ DB is stored in async_eq */ + struct mlx5_nb cq_err_nb; + struct mutex lock; /* sync async eqs creations */ int num_comp_vectors; struct mlx5_irq_info *irq_info; @@ -235,20 +238,6 @@ static struct mlx5_core_cq *mlx5_eq_cq_get(struct mlx5_eq *eq, u32 cqn) return cq; } -static void mlx5_eq_cq_event(struct mlx5_eq *eq, u32 cqn, int event_type) -{ - struct mlx5_core_cq *cq = mlx5_eq_cq_get(eq, cqn); - - if (unlikely(!cq)) { - mlx5_core_warn(eq->dev, "Async event for bogus CQ 0x%x\n", cqn); - return; - } - - cq->event(cq, event_type); - - mlx5_cq_put(cq); -} - static irqreturn_t mlx5_eq_comp_int(int irq, void *eq_ptr) { struct mlx5_eq_comp *eq_comp = eq_ptr; @@ -323,7 +312,6 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) struct mlx5_core_dev *dev; struct mlx5_eqe *eqe; int set_ci = 0; - u32 cqn = -1; u8 port; dev = eq->dev; @@ -358,12 +346,6 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) port, eqe->sub_type); } break; - case MLX5_EVENT_TYPE_CQ_ERROR: - cqn = be32_to_cpu(eqe->data.cq_err.cqn) & 0xffffff; - mlx5_core_warn(dev, "CQ error on CQN 0x%x, syndrome 0x%x\n", - cqn, eqe->data.cq_err.syndrome); - mlx5_eq_cq_event(eq, cqn, eqe->type); - break; case MLX5_EVENT_TYPE_PORT_MODULE_EVENT: mlx5_port_module_event(dev, eqe); @@ -639,6 +621,38 @@ static int destroy_async_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq) return err; } +static int cq_err_event_notifier(struct notifier_block *nb, + unsigned long type, void *data) +{ + struct mlx5_eq_table *eqt; + struct mlx5_core_cq *cq; + struct mlx5_eqe *eqe; + struct mlx5_eq *eq; + u32 cqn; + + /* type == MLX5_EVENT_TYPE_CQ_ERROR */ + + eqt = mlx5_nb_cof(nb, struct mlx5_eq_table, cq_err_nb); + eq = &eqt->async_eq; + eqe = data; + + cqn = be32_to_cpu(eqe->data.cq_err.cqn) & 0xffffff; + mlx5_core_warn(eq->dev, "CQ error on CQN 0x%x, syndrome 0x%x\n", + cqn, eqe->data.cq_err.syndrome); + + cq = mlx5_eq_cq_get(eq, cqn); + if (unlikely(!cq)) { + mlx5_core_warn(eq->dev, "Async event for bogus CQ 0x%x\n", cqn); + return NOTIFY_OK; + } + + cq->event(cq, type); + + mlx5_cq_put(cq); + + return NOTIFY_OK; +} + static u64 gather_async_events_mask(struct mlx5_core_dev *dev) { u64 async_event_mask = MLX5_ASYNC_EVENT_MASK; @@ -679,6 +693,9 @@ static int create_async_eqs(struct mlx5_core_dev *dev) struct mlx5_eq_param param = {}; int err; + MLX5_NB_INIT(&table->cq_err_nb, cq_err_event_notifier, CQ_ERROR); + mlx5_eq_notifier_register(dev, &table->cq_err_nb); + param = (struct mlx5_eq_param) { .index = MLX5_EQ_CMD_IDX, .mask = 1ull << MLX5_EVENT_TYPE_CMD, @@ -689,7 +706,7 @@ static int create_async_eqs(struct mlx5_core_dev *dev) err = create_async_eq(dev, "mlx5_cmd_eq", &table->cmd_eq, ¶m); if (err) { mlx5_core_warn(dev, "failed to create cmd EQ %d\n", err); - return err; + goto err0; } mlx5_cmd_use_events(dev); @@ -728,6 +745,8 @@ err2: err1: mlx5_cmd_use_polling(dev); destroy_async_eq(dev, &table->cmd_eq); +err0: + mlx5_eq_notifier_unregister(dev, &table->cq_err_nb); return err; } @@ -745,12 +764,15 @@ static void destroy_async_eqs(struct mlx5_core_dev *dev) if (err) mlx5_core_err(dev, "failed to destroy async eq, err(%d)\n", err); + mlx5_cmd_use_polling(dev); err = destroy_async_eq(dev, &table->cmd_eq); if (err) mlx5_core_err(dev, "failed to destroy command eq, err(%d)\n", err); + + mlx5_eq_notifier_unregister(dev, &table->cq_err_nb); } struct mlx5_eq *mlx5_get_async_eq(struct mlx5_core_dev *dev) -- cgit v1.2.3-55-g7522 From 69c1280b1f3b9123bc5154b2062507abcc14c3ef Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Tue, 20 Nov 2018 14:12:27 -0800 Subject: net/mlx5: Device events, Use async events chain Move all the generic async events handling into new specific events handling file events.c to keep eq.c file clean from concrete event logic handling. Use new API to register for NOTIFY_ANY to handle generic events and dispatch allowed events to mlx5_core consumers (mlx5_ib and mlx5e) Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/Makefile | 2 +- drivers/net/ethernet/mellanox/mlx5/core/en_stats.c | 9 +- drivers/net/ethernet/mellanox/mlx5/core/eq.c | 157 ------------ drivers/net/ethernet/mellanox/mlx5/core/events.c | 283 +++++++++++++++++++++ drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h | 34 +++ drivers/net/ethernet/mellanox/mlx5/core/main.c | 16 +- .../net/ethernet/mellanox/mlx5/core/mlx5_core.h | 6 +- drivers/net/ethernet/mellanox/mlx5/core/port.c | 57 ----- include/linux/mlx5/driver.h | 29 +-- include/linux/mlx5/port.h | 3 - 10 files changed, 344 insertions(+), 252 deletions(-) create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/events.c (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile index d324a3884462..26afe0779a0c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile +++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile @@ -14,7 +14,7 @@ obj-$(CONFIG_MLX5_CORE) += mlx5_core.o mlx5_core-y := main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \ health.o mcg.o cq.o srq.o alloc.o qp.o port.o mr.o pd.o \ mad.o transobj.o vport.o sriov.o fs_cmd.o fs_core.o \ - fs_counters.o rl.o lag.o dev.o wq.o lib/gid.o \ + fs_counters.o rl.o lag.o dev.o events.o wq.o lib/gid.o \ diag/fs_tracepoint.o diag/fw_tracer.o # diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c index 1e55b9c27ffc..748d23806391 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c @@ -30,6 +30,7 @@ * SOFTWARE. */ +#include "lib/mlx5.h" #include "en.h" #include "en_accel/ipsec.h" #include "en_accel/tls.h" @@ -1120,15 +1121,17 @@ static int mlx5e_grp_pme_fill_strings(struct mlx5e_priv *priv, u8 *data, static int mlx5e_grp_pme_fill_stats(struct mlx5e_priv *priv, u64 *data, int idx) { - struct mlx5_priv *mlx5_priv = &priv->mdev->priv; + struct mlx5_pme_stats pme_stats; int i; + mlx5_get_pme_stats(priv->mdev, &pme_stats); + for (i = 0; i < NUM_PME_STATUS_STATS; i++) - data[idx++] = MLX5E_READ_CTR64_CPU(mlx5_priv->pme_stats.status_counters, + data[idx++] = MLX5E_READ_CTR64_CPU(pme_stats.status_counters, mlx5e_pme_status_desc, i); for (i = 0; i < NUM_PME_ERR_STATS; i++) - data[idx++] = MLX5E_READ_CTR64_CPU(mlx5_priv->pme_stats.error_counters, + data[idx++] = MLX5E_READ_CTR64_CPU(pme_stats.error_counters, mlx5e_pme_error_desc, i); return idx; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 4e3febbf639d..4aa39a1fe23f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -108,121 +108,6 @@ static int mlx5_cmd_destroy_eq(struct mlx5_core_dev *dev, u8 eqn) return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); } -static const char *eqe_type_str(u8 type) -{ - switch (type) { - case MLX5_EVENT_TYPE_COMP: - return "MLX5_EVENT_TYPE_COMP"; - case MLX5_EVENT_TYPE_PATH_MIG: - return "MLX5_EVENT_TYPE_PATH_MIG"; - case MLX5_EVENT_TYPE_COMM_EST: - return "MLX5_EVENT_TYPE_COMM_EST"; - case MLX5_EVENT_TYPE_SQ_DRAINED: - return "MLX5_EVENT_TYPE_SQ_DRAINED"; - case MLX5_EVENT_TYPE_SRQ_LAST_WQE: - return "MLX5_EVENT_TYPE_SRQ_LAST_WQE"; - case MLX5_EVENT_TYPE_SRQ_RQ_LIMIT: - return "MLX5_EVENT_TYPE_SRQ_RQ_LIMIT"; - case MLX5_EVENT_TYPE_CQ_ERROR: - return "MLX5_EVENT_TYPE_CQ_ERROR"; - case MLX5_EVENT_TYPE_WQ_CATAS_ERROR: - return "MLX5_EVENT_TYPE_WQ_CATAS_ERROR"; - case MLX5_EVENT_TYPE_PATH_MIG_FAILED: - return "MLX5_EVENT_TYPE_PATH_MIG_FAILED"; - case MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR: - return "MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR"; - case MLX5_EVENT_TYPE_WQ_ACCESS_ERROR: - return "MLX5_EVENT_TYPE_WQ_ACCESS_ERROR"; - case MLX5_EVENT_TYPE_SRQ_CATAS_ERROR: - return "MLX5_EVENT_TYPE_SRQ_CATAS_ERROR"; - case MLX5_EVENT_TYPE_INTERNAL_ERROR: - return "MLX5_EVENT_TYPE_INTERNAL_ERROR"; - case MLX5_EVENT_TYPE_PORT_CHANGE: - return "MLX5_EVENT_TYPE_PORT_CHANGE"; - case MLX5_EVENT_TYPE_GPIO_EVENT: - return "MLX5_EVENT_TYPE_GPIO_EVENT"; - case MLX5_EVENT_TYPE_PORT_MODULE_EVENT: - return "MLX5_EVENT_TYPE_PORT_MODULE_EVENT"; - case MLX5_EVENT_TYPE_TEMP_WARN_EVENT: - return "MLX5_EVENT_TYPE_TEMP_WARN_EVENT"; - case MLX5_EVENT_TYPE_REMOTE_CONFIG: - return "MLX5_EVENT_TYPE_REMOTE_CONFIG"; - case MLX5_EVENT_TYPE_DB_BF_CONGESTION: - return "MLX5_EVENT_TYPE_DB_BF_CONGESTION"; - case MLX5_EVENT_TYPE_STALL_EVENT: - return "MLX5_EVENT_TYPE_STALL_EVENT"; - case MLX5_EVENT_TYPE_CMD: - return "MLX5_EVENT_TYPE_CMD"; - case MLX5_EVENT_TYPE_PAGE_REQUEST: - return "MLX5_EVENT_TYPE_PAGE_REQUEST"; - case MLX5_EVENT_TYPE_PAGE_FAULT: - return "MLX5_EVENT_TYPE_PAGE_FAULT"; - case MLX5_EVENT_TYPE_PPS_EVENT: - return "MLX5_EVENT_TYPE_PPS_EVENT"; - case MLX5_EVENT_TYPE_NIC_VPORT_CHANGE: - return "MLX5_EVENT_TYPE_NIC_VPORT_CHANGE"; - case MLX5_EVENT_TYPE_FPGA_ERROR: - return "MLX5_EVENT_TYPE_FPGA_ERROR"; - case MLX5_EVENT_TYPE_FPGA_QP_ERROR: - return "MLX5_EVENT_TYPE_FPGA_QP_ERROR"; - case MLX5_EVENT_TYPE_GENERAL_EVENT: - return "MLX5_EVENT_TYPE_GENERAL_EVENT"; - case MLX5_EVENT_TYPE_DEVICE_TRACER: - return "MLX5_EVENT_TYPE_DEVICE_TRACER"; - default: - return "Unrecognized event"; - } -} - -static enum mlx5_dev_event port_subtype_event(u8 subtype) -{ - switch (subtype) { - case MLX5_PORT_CHANGE_SUBTYPE_DOWN: - return MLX5_DEV_EVENT_PORT_DOWN; - case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE: - return MLX5_DEV_EVENT_PORT_UP; - case MLX5_PORT_CHANGE_SUBTYPE_INITIALIZED: - return MLX5_DEV_EVENT_PORT_INITIALIZED; - case MLX5_PORT_CHANGE_SUBTYPE_LID: - return MLX5_DEV_EVENT_LID_CHANGE; - case MLX5_PORT_CHANGE_SUBTYPE_PKEY: - return MLX5_DEV_EVENT_PKEY_CHANGE; - case MLX5_PORT_CHANGE_SUBTYPE_GUID: - return MLX5_DEV_EVENT_GUID_CHANGE; - case MLX5_PORT_CHANGE_SUBTYPE_CLIENT_REREG: - return MLX5_DEV_EVENT_CLIENT_REREG; - } - return -1; -} - -static void general_event_handler(struct mlx5_core_dev *dev, - struct mlx5_eqe *eqe) -{ - switch (eqe->sub_type) { - case MLX5_GENERAL_SUBTYPE_DELAY_DROP_TIMEOUT: - if (dev->event) - dev->event(dev, MLX5_DEV_EVENT_DELAY_DROP_TIMEOUT, 0); - break; - default: - mlx5_core_dbg(dev, "General event with unrecognized subtype: sub_type %d\n", - eqe->sub_type); - } -} - -static void mlx5_temp_warning_event(struct mlx5_core_dev *dev, - struct mlx5_eqe *eqe) -{ - u64 value_lsb; - u64 value_msb; - - value_lsb = be64_to_cpu(eqe->data.temp_warning.sensor_warning_lsb); - value_msb = be64_to_cpu(eqe->data.temp_warning.sensor_warning_msb); - - mlx5_core_warn(dev, - "High temperature on sensors with bit set %llx %llx", - value_msb, value_lsb); -} - /* caller must eventually call mlx5_cq_put on the returned cq */ static struct mlx5_core_cq *mlx5_eq_cq_get(struct mlx5_eq *eq, u32 cqn) { @@ -312,7 +197,6 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) struct mlx5_core_dev *dev; struct mlx5_eqe *eqe; int set_ci = 0; - u8 port; dev = eq->dev; eqt = dev->priv.eq_table; @@ -324,47 +208,6 @@ static irqreturn_t mlx5_eq_async_int(int irq, void *eq_ptr) */ dma_rmb(); - mlx5_core_dbg(eq->dev, "eqn %d, eqe type %s\n", - eq->eqn, eqe_type_str(eqe->type)); - switch (eqe->type) { - case MLX5_EVENT_TYPE_PORT_CHANGE: - port = (eqe->data.port.port >> 4) & 0xf; - switch (eqe->sub_type) { - case MLX5_PORT_CHANGE_SUBTYPE_DOWN: - case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE: - case MLX5_PORT_CHANGE_SUBTYPE_LID: - case MLX5_PORT_CHANGE_SUBTYPE_PKEY: - case MLX5_PORT_CHANGE_SUBTYPE_GUID: - case MLX5_PORT_CHANGE_SUBTYPE_CLIENT_REREG: - case MLX5_PORT_CHANGE_SUBTYPE_INITIALIZED: - if (dev->event) - dev->event(dev, port_subtype_event(eqe->sub_type), - (unsigned long)port); - break; - default: - mlx5_core_warn(dev, "Port event with unrecognized subtype: port %d, sub_type %d\n", - port, eqe->sub_type); - } - break; - - case MLX5_EVENT_TYPE_PORT_MODULE_EVENT: - mlx5_port_module_event(dev, eqe); - break; - - case MLX5_EVENT_TYPE_TEMP_WARN_EVENT: - mlx5_temp_warning_event(dev, eqe); - break; - - case MLX5_EVENT_TYPE_GENERAL_EVENT: - general_event_handler(dev, eqe); - break; - - default: - mlx5_core_warn(dev, "Unhandled event 0x%x on EQ 0x%x\n", - eqe->type, eq->eqn); - break; - } - if (likely(eqe->type < MLX5_EVENT_TYPE_MAX)) atomic_notifier_call_chain(&eqt->nh[eqe->type], eqe->type, eqe); else diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c new file mode 100644 index 000000000000..d3ab86bd394b --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -0,0 +1,283 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB +// Copyright (c) 2018 Mellanox Technologies + +#include +#include "mlx5_core.h" +#include "lib/eq.h" +#include "lib/mlx5.h" + +struct mlx5_events { + struct mlx5_nb nb; + struct mlx5_core_dev *dev; + + /* port module evetns stats */ + struct mlx5_pme_stats pme_stats; +}; + +static const char *eqe_type_str(u8 type) +{ + switch (type) { + case MLX5_EVENT_TYPE_COMP: + return "MLX5_EVENT_TYPE_COMP"; + case MLX5_EVENT_TYPE_PATH_MIG: + return "MLX5_EVENT_TYPE_PATH_MIG"; + case MLX5_EVENT_TYPE_COMM_EST: + return "MLX5_EVENT_TYPE_COMM_EST"; + case MLX5_EVENT_TYPE_SQ_DRAINED: + return "MLX5_EVENT_TYPE_SQ_DRAINED"; + case MLX5_EVENT_TYPE_SRQ_LAST_WQE: + return "MLX5_EVENT_TYPE_SRQ_LAST_WQE"; + case MLX5_EVENT_TYPE_SRQ_RQ_LIMIT: + return "MLX5_EVENT_TYPE_SRQ_RQ_LIMIT"; + case MLX5_EVENT_TYPE_CQ_ERROR: + return "MLX5_EVENT_TYPE_CQ_ERROR"; + case MLX5_EVENT_TYPE_WQ_CATAS_ERROR: + return "MLX5_EVENT_TYPE_WQ_CATAS_ERROR"; + case MLX5_EVENT_TYPE_PATH_MIG_FAILED: + return "MLX5_EVENT_TYPE_PATH_MIG_FAILED"; + case MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR: + return "MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR"; + case MLX5_EVENT_TYPE_WQ_ACCESS_ERROR: + return "MLX5_EVENT_TYPE_WQ_ACCESS_ERROR"; + case MLX5_EVENT_TYPE_SRQ_CATAS_ERROR: + return "MLX5_EVENT_TYPE_SRQ_CATAS_ERROR"; + case MLX5_EVENT_TYPE_INTERNAL_ERROR: + return "MLX5_EVENT_TYPE_INTERNAL_ERROR"; + case MLX5_EVENT_TYPE_PORT_CHANGE: + return "MLX5_EVENT_TYPE_PORT_CHANGE"; + case MLX5_EVENT_TYPE_GPIO_EVENT: + return "MLX5_EVENT_TYPE_GPIO_EVENT"; + case MLX5_EVENT_TYPE_PORT_MODULE_EVENT: + return "MLX5_EVENT_TYPE_PORT_MODULE_EVENT"; + case MLX5_EVENT_TYPE_TEMP_WARN_EVENT: + return "MLX5_EVENT_TYPE_TEMP_WARN_EVENT"; + case MLX5_EVENT_TYPE_REMOTE_CONFIG: + return "MLX5_EVENT_TYPE_REMOTE_CONFIG"; + case MLX5_EVENT_TYPE_DB_BF_CONGESTION: + return "MLX5_EVENT_TYPE_DB_BF_CONGESTION"; + case MLX5_EVENT_TYPE_STALL_EVENT: + return "MLX5_EVENT_TYPE_STALL_EVENT"; + case MLX5_EVENT_TYPE_CMD: + return "MLX5_EVENT_TYPE_CMD"; + case MLX5_EVENT_TYPE_PAGE_REQUEST: + return "MLX5_EVENT_TYPE_PAGE_REQUEST"; + case MLX5_EVENT_TYPE_PAGE_FAULT: + return "MLX5_EVENT_TYPE_PAGE_FAULT"; + case MLX5_EVENT_TYPE_PPS_EVENT: + return "MLX5_EVENT_TYPE_PPS_EVENT"; + case MLX5_EVENT_TYPE_NIC_VPORT_CHANGE: + return "MLX5_EVENT_TYPE_NIC_VPORT_CHANGE"; + case MLX5_EVENT_TYPE_FPGA_ERROR: + return "MLX5_EVENT_TYPE_FPGA_ERROR"; + case MLX5_EVENT_TYPE_FPGA_QP_ERROR: + return "MLX5_EVENT_TYPE_FPGA_QP_ERROR"; + case MLX5_EVENT_TYPE_GENERAL_EVENT: + return "MLX5_EVENT_TYPE_GENERAL_EVENT"; + case MLX5_EVENT_TYPE_DEVICE_TRACER: + return "MLX5_EVENT_TYPE_DEVICE_TRACER"; + default: + return "Unrecognized event"; + } +} + +static enum mlx5_dev_event port_subtype2dev(u8 subtype) +{ + switch (subtype) { + case MLX5_PORT_CHANGE_SUBTYPE_DOWN: + return MLX5_DEV_EVENT_PORT_DOWN; + case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE: + return MLX5_DEV_EVENT_PORT_UP; + case MLX5_PORT_CHANGE_SUBTYPE_INITIALIZED: + return MLX5_DEV_EVENT_PORT_INITIALIZED; + case MLX5_PORT_CHANGE_SUBTYPE_LID: + return MLX5_DEV_EVENT_LID_CHANGE; + case MLX5_PORT_CHANGE_SUBTYPE_PKEY: + return MLX5_DEV_EVENT_PKEY_CHANGE; + case MLX5_PORT_CHANGE_SUBTYPE_GUID: + return MLX5_DEV_EVENT_GUID_CHANGE; + case MLX5_PORT_CHANGE_SUBTYPE_CLIENT_REREG: + return MLX5_DEV_EVENT_CLIENT_REREG; + } + return -1; +} + +static void temp_warning_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe) +{ + u64 value_lsb; + u64 value_msb; + + value_lsb = be64_to_cpu(eqe->data.temp_warning.sensor_warning_lsb); + value_msb = be64_to_cpu(eqe->data.temp_warning.sensor_warning_msb); + + mlx5_core_warn(dev, + "High temperature on sensors with bit set %llx %llx", + value_msb, value_lsb); +} + +static const char *mlx5_pme_status[MLX5_MODULE_STATUS_NUM] = { + "Cable plugged", /* MLX5_MODULE_STATUS_PLUGGED = 0x1 */ + "Cable unplugged", /* MLX5_MODULE_STATUS_UNPLUGGED = 0x2 */ + "Cable error", /* MLX5_MODULE_STATUS_ERROR = 0x3 */ +}; + +static const char *mlx5_pme_error[MLX5_MODULE_EVENT_ERROR_NUM] = { + "Power budget exceeded", + "Long Range for non MLNX cable", + "Bus stuck(I2C or data shorted)", + "No EEPROM/retry timeout", + "Enforce part number list", + "Unknown identifier", + "High Temperature", + "Bad or shorted cable/module", + "Unknown status", +}; + +static void port_module_event(struct mlx5_events *events, struct mlx5_eqe *eqe) +{ + enum port_module_event_status_type module_status; + enum port_module_event_error_type error_type; + struct mlx5_eqe_port_module *module_event_eqe; + struct mlx5_core_dev *dev = events->dev; + u8 module_num; + + module_event_eqe = &eqe->data.port_module; + module_num = module_event_eqe->module; + module_status = module_event_eqe->module_status & + PORT_MODULE_EVENT_MODULE_STATUS_MASK; + error_type = module_event_eqe->error_type & + PORT_MODULE_EVENT_ERROR_TYPE_MASK; + + if (module_status < MLX5_MODULE_STATUS_ERROR) { + events->pme_stats.status_counters[module_status - 1]++; + } else if (module_status == MLX5_MODULE_STATUS_ERROR) { + if (error_type >= MLX5_MODULE_EVENT_ERROR_UNKNOWN) + /* Unknown error type */ + error_type = MLX5_MODULE_EVENT_ERROR_UNKNOWN; + events->pme_stats.error_counters[error_type]++; + } + + if (!printk_ratelimit()) + return; + + if (module_status < MLX5_MODULE_STATUS_ERROR) + mlx5_core_info(dev, + "Port module event: module %u, %s\n", + module_num, mlx5_pme_status[module_status - 1]); + + else if (module_status == MLX5_MODULE_STATUS_ERROR) + mlx5_core_info(dev, + "Port module event[error]: module %u, %s, %s\n", + module_num, mlx5_pme_status[module_status - 1], + mlx5_pme_error[error_type]); +} + +void mlx5_get_pme_stats(struct mlx5_core_dev *dev, struct mlx5_pme_stats *stats) +{ + *stats = dev->priv.events->pme_stats; +} + +/* Event handler for the low level mlx5_core driver. + * This handler will process/filter _some_ events and sometimes dispatch + * the equivalent mlx5_dev_event to the HCA interfaces (mlx5_ib and mlx5e) + * + * Other Major feature specific events such as + * clock/eswitch/fpga/FW trace and many others, are handled elsewhere, with + * separate notifiers callbacks, specifically by those mlx5 components. + */ +static int events_notifier(struct notifier_block *nb, + unsigned long type, void *data) +{ + bool dev_event_dispatch = false; + enum mlx5_dev_event dev_event; + unsigned long dev_event_data; + + struct mlx5_eqe *eqe = data; + struct mlx5_events *events; + struct mlx5_core_dev *dev; + u8 port; + + events = mlx5_nb_cof(nb, struct mlx5_events, nb); + dev = events->dev; + + mlx5_core_dbg(dev, "Async eqe type %s, subtype (%d)\n", + eqe_type_str(eqe->type), eqe->sub_type); + switch (eqe->type) { + case MLX5_EVENT_TYPE_PORT_CHANGE: + port = (eqe->data.port.port >> 4) & 0xf; + switch (eqe->sub_type) { + case MLX5_PORT_CHANGE_SUBTYPE_DOWN: + case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE: + case MLX5_PORT_CHANGE_SUBTYPE_LID: + case MLX5_PORT_CHANGE_SUBTYPE_PKEY: + case MLX5_PORT_CHANGE_SUBTYPE_GUID: + case MLX5_PORT_CHANGE_SUBTYPE_CLIENT_REREG: + case MLX5_PORT_CHANGE_SUBTYPE_INITIALIZED: + dev_event = port_subtype2dev(eqe->sub_type); + dev_event_data = (unsigned long)port; + dev_event_dispatch = true; + break; + default: + mlx5_core_warn(dev, "Port event with unrecognized subtype: port %d, sub_type %d\n", + port, eqe->sub_type); + } + break; + case MLX5_EVENT_TYPE_GENERAL_EVENT: + switch (eqe->sub_type) { + case MLX5_GENERAL_SUBTYPE_DELAY_DROP_TIMEOUT: + dev_event = MLX5_DEV_EVENT_DELAY_DROP_TIMEOUT; + dev_event_data = 0; + dev_event_dispatch = true; + break; + default: + mlx5_core_dbg(dev, "General event with unrecognized subtype: sub_type %d\n", + eqe->sub_type); + } + break; + + case MLX5_EVENT_TYPE_PORT_MODULE_EVENT: + port_module_event(events, eqe); + break; + case MLX5_EVENT_TYPE_TEMP_WARN_EVENT: + temp_warning_event(dev, eqe); + break; + default: + return NOTIFY_DONE; + } + + if (dev->event && dev_event_dispatch) + dev->event(dev, dev_event, dev_event_data); + + return NOTIFY_OK; +} + +int mlx5_events_init(struct mlx5_core_dev *dev) +{ + struct mlx5_events *events = kzalloc(sizeof(*events), GFP_KERNEL); + + if (!events) + return -ENOMEM; + + events->dev = dev; + dev->priv.events = events; + return 0; +} + +void mlx5_events_cleanup(struct mlx5_core_dev *dev) +{ + kvfree(dev->priv.events); +} + +void mlx5_events_start(struct mlx5_core_dev *dev) +{ + struct mlx5_events *events = dev->priv.events; + + MLX5_NB_INIT(&events->nb, events_notifier, NOTIFY_ANY); + mlx5_eq_notifier_register(dev, &events->nb); +} + +void mlx5_events_stop(struct mlx5_core_dev *dev) +{ + struct mlx5_events *events = dev->priv.events; + + mlx5_eq_notifier_unregister(dev, &events->nb); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h index 7550b1cc8c6a..23317e328b0b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h @@ -33,6 +33,8 @@ #ifndef __LIB_MLX5_H__ #define __LIB_MLX5_H__ +#include "mlx5_core.h" + void mlx5_init_reserved_gids(struct mlx5_core_dev *dev); void mlx5_cleanup_reserved_gids(struct mlx5_core_dev *dev); int mlx5_core_reserve_gids(struct mlx5_core_dev *dev, unsigned int count); @@ -40,4 +42,36 @@ void mlx5_core_unreserve_gids(struct mlx5_core_dev *dev, unsigned int count); int mlx5_core_reserved_gid_alloc(struct mlx5_core_dev *dev, int *gid_index); void mlx5_core_reserved_gid_free(struct mlx5_core_dev *dev, int gid_index); +/* TODO move to lib/events.h */ + +#define PORT_MODULE_EVENT_MODULE_STATUS_MASK 0xF +#define PORT_MODULE_EVENT_ERROR_TYPE_MASK 0xF + +enum port_module_event_status_type { + MLX5_MODULE_STATUS_PLUGGED = 0x1, + MLX5_MODULE_STATUS_UNPLUGGED = 0x2, + MLX5_MODULE_STATUS_ERROR = 0x3, + MLX5_MODULE_STATUS_NUM = 0x3, +}; + +enum port_module_event_error_type { + MLX5_MODULE_EVENT_ERROR_POWER_BUDGET_EXCEEDED, + MLX5_MODULE_EVENT_ERROR_LONG_RANGE_FOR_NON_MLNX_CABLE_MODULE, + MLX5_MODULE_EVENT_ERROR_BUS_STUCK, + MLX5_MODULE_EVENT_ERROR_NO_EEPROM_RETRY_TIMEOUT, + MLX5_MODULE_EVENT_ERROR_ENFORCE_PART_NUMBER_LIST, + MLX5_MODULE_EVENT_ERROR_UNKNOWN_IDENTIFIER, + MLX5_MODULE_EVENT_ERROR_HIGH_TEMPERATURE, + MLX5_MODULE_EVENT_ERROR_BAD_CABLE, + MLX5_MODULE_EVENT_ERROR_UNKNOWN, + MLX5_MODULE_EVENT_ERROR_NUM, +}; + +struct mlx5_pme_stats { + u64 status_counters[MLX5_MODULE_STATUS_NUM]; + u64 error_counters[MLX5_MODULE_EVENT_ERROR_NUM]; +}; + +void mlx5_get_pme_stats(struct mlx5_core_dev *dev, struct mlx5_pme_stats *stats); + #endif diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 9e4cd2757ea8..e56278ead4eb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -735,10 +735,16 @@ static int mlx5_init_once(struct mlx5_core_dev *dev, struct mlx5_priv *priv) goto out; } + err = mlx5_events_init(dev); + if (err) { + dev_err(&pdev->dev, "failed to initialize events\n"); + goto err_eq_cleanup; + } + err = mlx5_cq_debugfs_init(dev); if (err) { dev_err(&pdev->dev, "failed to initialize cq debugfs\n"); - goto err_eq_cleanup; + goto err_events_cleanup; } mlx5_init_qp_table(dev); @@ -801,7 +807,8 @@ err_tables_cleanup: mlx5_cleanup_srq_table(dev); mlx5_cleanup_qp_table(dev); mlx5_cq_debugfs_cleanup(dev); - +err_events_cleanup: + mlx5_events_cleanup(dev); err_eq_cleanup: mlx5_eq_table_cleanup(dev); @@ -824,6 +831,7 @@ static void mlx5_cleanup_once(struct mlx5_core_dev *dev) mlx5_cleanup_srq_table(dev); mlx5_cleanup_qp_table(dev); mlx5_cq_debugfs_cleanup(dev); + mlx5_events_cleanup(dev); mlx5_eq_table_cleanup(dev); } @@ -947,6 +955,7 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, goto err_get_uars; } + mlx5_events_start(dev); mlx5_pagealloc_start(dev); err = mlx5_eq_table_create(dev); @@ -1036,6 +1045,7 @@ err_fw_tracer: err_eq_table: mlx5_pagealloc_stop(dev); + mlx5_events_stop(dev); mlx5_put_uars_page(dev, priv->uar); err_get_uars: @@ -1095,8 +1105,8 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, mlx5_fw_tracer_cleanup(dev->tracer); mlx5_eq_table_destroy(dev); mlx5_pagealloc_stop(dev); + mlx5_events_stop(dev); mlx5_put_uars_page(dev, priv->uar); - if (cleanup) mlx5_cleanup_once(dev); mlx5_stop_health_poll(dev, cleanup); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index 5dd453e47a04..c70bd94e18d6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -105,7 +105,6 @@ int mlx5_cmd_fast_teardown_hca(struct mlx5_core_dev *dev); void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event, unsigned long param); -void mlx5_port_module_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe); void mlx5_enter_error_state(struct mlx5_core_dev *dev, bool force); void mlx5_disable_device(struct mlx5_core_dev *dev); void mlx5_recover_device(struct mlx5_core_dev *dev); @@ -141,6 +140,11 @@ int mlx5_query_qcam_reg(struct mlx5_core_dev *mdev, u32 *qcam, void mlx5_lag_add(struct mlx5_core_dev *dev, struct net_device *netdev); void mlx5_lag_remove(struct mlx5_core_dev *dev); +int mlx5_events_init(struct mlx5_core_dev *dev); +void mlx5_events_cleanup(struct mlx5_core_dev *dev); +void mlx5_events_start(struct mlx5_core_dev *dev); +void mlx5_events_stop(struct mlx5_core_dev *dev); + void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv); void mlx5_remove_device(struct mlx5_interface *intf, struct mlx5_priv *priv); void mlx5_attach_device(struct mlx5_core_dev *dev); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/port.c b/drivers/net/ethernet/mellanox/mlx5/core/port.c index 31a9cbd85689..2b82f35f4c35 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/port.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/port.c @@ -915,63 +915,6 @@ void mlx5_query_port_fcs(struct mlx5_core_dev *mdev, bool *supported, *enabled = !!(MLX5_GET(pcmr_reg, out, fcs_chk)); } -static const char *mlx5_pme_status[MLX5_MODULE_STATUS_NUM] = { - "Cable plugged", /* MLX5_MODULE_STATUS_PLUGGED = 0x1 */ - "Cable unplugged", /* MLX5_MODULE_STATUS_UNPLUGGED = 0x2 */ - "Cable error", /* MLX5_MODULE_STATUS_ERROR = 0x3 */ -}; - -static const char *mlx5_pme_error[MLX5_MODULE_EVENT_ERROR_NUM] = { - "Power budget exceeded", - "Long Range for non MLNX cable", - "Bus stuck(I2C or data shorted)", - "No EEPROM/retry timeout", - "Enforce part number list", - "Unknown identifier", - "High Temperature", - "Bad or shorted cable/module", - "Unknown status", -}; - -void mlx5_port_module_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe) -{ - enum port_module_event_status_type module_status; - enum port_module_event_error_type error_type; - struct mlx5_eqe_port_module *module_event_eqe; - struct mlx5_priv *priv = &dev->priv; - u8 module_num; - - module_event_eqe = &eqe->data.port_module; - module_num = module_event_eqe->module; - module_status = module_event_eqe->module_status & - PORT_MODULE_EVENT_MODULE_STATUS_MASK; - error_type = module_event_eqe->error_type & - PORT_MODULE_EVENT_ERROR_TYPE_MASK; - - if (module_status < MLX5_MODULE_STATUS_ERROR) { - priv->pme_stats.status_counters[module_status - 1]++; - } else if (module_status == MLX5_MODULE_STATUS_ERROR) { - if (error_type >= MLX5_MODULE_EVENT_ERROR_UNKNOWN) - /* Unknown error type */ - error_type = MLX5_MODULE_EVENT_ERROR_UNKNOWN; - priv->pme_stats.error_counters[error_type]++; - } - - if (!printk_ratelimit()) - return; - - if (module_status < MLX5_MODULE_STATUS_ERROR) - mlx5_core_info(dev, - "Port module event: module %u, %s\n", - module_num, mlx5_pme_status[module_status - 1]); - - else if (module_status == MLX5_MODULE_STATUS_ERROR) - mlx5_core_info(dev, - "Port module event[error]: module %u, %s, %s\n", - module_num, mlx5_pme_status[module_status - 1], - mlx5_pme_error[error_type]); -} - int mlx5_query_mtpps(struct mlx5_core_dev *mdev, u32 *mtpps, u32 mtpps_size) { u32 in[MLX5_ST_SZ_DW(mtpps_reg)] = {0}; diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index afba0864f45c..ba64ecf72478 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -514,6 +514,7 @@ struct mlx5_fc_stats { unsigned long sampling_interval; /* jiffies */ }; +struct mlx5_events; struct mlx5_mpfs; struct mlx5_eswitch; struct mlx5_lag; @@ -540,31 +541,6 @@ struct mlx5_rl_table { struct mlx5_rl_entry *rl_entry; }; -enum port_module_event_status_type { - MLX5_MODULE_STATUS_PLUGGED = 0x1, - MLX5_MODULE_STATUS_UNPLUGGED = 0x2, - MLX5_MODULE_STATUS_ERROR = 0x3, - MLX5_MODULE_STATUS_NUM = 0x3, -}; - -enum port_module_event_error_type { - MLX5_MODULE_EVENT_ERROR_POWER_BUDGET_EXCEEDED, - MLX5_MODULE_EVENT_ERROR_LONG_RANGE_FOR_NON_MLNX_CABLE_MODULE, - MLX5_MODULE_EVENT_ERROR_BUS_STUCK, - MLX5_MODULE_EVENT_ERROR_NO_EEPROM_RETRY_TIMEOUT, - MLX5_MODULE_EVENT_ERROR_ENFORCE_PART_NUMBER_LIST, - MLX5_MODULE_EVENT_ERROR_UNKNOWN_IDENTIFIER, - MLX5_MODULE_EVENT_ERROR_HIGH_TEMPERATURE, - MLX5_MODULE_EVENT_ERROR_BAD_CABLE, - MLX5_MODULE_EVENT_ERROR_UNKNOWN, - MLX5_MODULE_EVENT_ERROR_NUM, -}; - -struct mlx5_port_module_event_stats { - u64 status_counters[MLX5_MODULE_STATUS_NUM]; - u64 error_counters[MLX5_MODULE_EVENT_ERROR_NUM]; -}; - struct mlx5_priv { char name[MLX5_MAX_NAME_LEN]; struct mlx5_eq_table *eq_table; @@ -614,6 +590,7 @@ struct mlx5_priv { struct list_head waiting_events_list; bool is_accum_events; + struct mlx5_events *events; struct mlx5_flow_steering *steering; struct mlx5_mpfs *mpfs; @@ -624,8 +601,6 @@ struct mlx5_priv { struct mlx5_fc_stats fc_stats; struct mlx5_rl_table rl_table; - struct mlx5_port_module_event_stats pme_stats; - struct mlx5_bfreg_data bfregs; struct mlx5_uars_page *uar; }; diff --git a/include/linux/mlx5/port.h b/include/linux/mlx5/port.h index 34aed6032f86..bf4bc01ffb0c 100644 --- a/include/linux/mlx5/port.h +++ b/include/linux/mlx5/port.h @@ -107,9 +107,6 @@ enum mlx5e_connector_type { #define MLX5E_PROT_MASK(link_mode) (1 << link_mode) -#define PORT_MODULE_EVENT_MODULE_STATUS_MASK 0xF -#define PORT_MODULE_EVENT_ERROR_TYPE_MASK 0xF - int mlx5_set_port_caps(struct mlx5_core_dev *dev, u8 port_num, u32 caps); int mlx5_query_port_ptys(struct mlx5_core_dev *dev, u32 *ptys, int ptys_size, int proto_mask, u8 local_port); -- cgit v1.2.3-55-g7522 From 2c89156082b3be7a064b09dfb0bcd274609f355d Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Tue, 20 Nov 2018 14:12:28 -0800 Subject: net/mlx5: Improve core device events handling Register a separate handler per event type, rather than listening for all events and looking for the events to handle in a switch case. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/events.c | 223 ++++++++++++++--------- 1 file changed, 136 insertions(+), 87 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index d3ab86bd394b..3ad004af37d7 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -2,15 +2,41 @@ // Copyright (c) 2018 Mellanox Technologies #include + #include "mlx5_core.h" #include "lib/eq.h" #include "lib/mlx5.h" +struct mlx5_event_nb { + struct mlx5_nb nb; + void *ctx; +}; + +/* General events handlers for the low level mlx5_core driver + * + * Other Major feature specific events such as + * clock/eswitch/fpga/FW trace and many others, are handled elsewhere, with + * separate notifiers callbacks, specifically by those mlx5 components. + */ +static int any_notifier(struct notifier_block *, unsigned long, void *); +static int port_change(struct notifier_block *, unsigned long, void *); +static int general_event(struct notifier_block *, unsigned long, void *); +static int temp_warn(struct notifier_block *, unsigned long, void *); +static int port_module(struct notifier_block *, unsigned long, void *); + +static struct mlx5_nb events_nbs_ref[] = { + {.nb.notifier_call = any_notifier, .event_type = MLX5_EVENT_TYPE_NOTIFY_ANY }, + {.nb.notifier_call = port_change, .event_type = MLX5_EVENT_TYPE_PORT_CHANGE }, + {.nb.notifier_call = general_event, .event_type = MLX5_EVENT_TYPE_GENERAL_EVENT }, + {.nb.notifier_call = temp_warn, .event_type = MLX5_EVENT_TYPE_TEMP_WARN_EVENT }, + {.nb.notifier_call = port_module, .event_type = MLX5_EVENT_TYPE_PORT_MODULE_EVENT }, +}; + struct mlx5_events { - struct mlx5_nb nb; struct mlx5_core_dev *dev; + struct mlx5_event_nb notifiers[ARRAY_SIZE(events_nbs_ref)]; - /* port module evetns stats */ + /* port module events stats */ struct mlx5_pme_stats pme_stats; }; @@ -80,6 +106,19 @@ static const char *eqe_type_str(u8 type) } } +/* handles all FW events, type == eqe->type */ +static int any_notifier(struct notifier_block *nb, + unsigned long type, void *data) +{ + struct mlx5_event_nb *event_nb = mlx5_nb_cof(nb, struct mlx5_event_nb, nb); + struct mlx5_events *events = event_nb->ctx; + struct mlx5_eqe *eqe = data; + + mlx5_core_dbg(events->dev, "Async eqe type %s, subtype (%d)\n", + eqe_type_str(eqe->type), eqe->sub_type); + return NOTIFY_OK; +} + static enum mlx5_dev_event port_subtype2dev(u8 subtype) { switch (subtype) { @@ -101,19 +140,92 @@ static enum mlx5_dev_event port_subtype2dev(u8 subtype) return -1; } -static void temp_warning_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe) +/* type == MLX5_EVENT_TYPE_PORT_CHANGE */ +static int port_change(struct notifier_block *nb, + unsigned long type, void *data) { + struct mlx5_event_nb *event_nb = mlx5_nb_cof(nb, struct mlx5_event_nb, nb); + struct mlx5_events *events = event_nb->ctx; + struct mlx5_core_dev *dev = events->dev; + + bool dev_event_dispatch = false; + enum mlx5_dev_event dev_event; + unsigned long dev_event_data; + struct mlx5_eqe *eqe = data; + u8 port = (eqe->data.port.port >> 4) & 0xf; + + switch (eqe->sub_type) { + case MLX5_PORT_CHANGE_SUBTYPE_DOWN: + case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE: + case MLX5_PORT_CHANGE_SUBTYPE_LID: + case MLX5_PORT_CHANGE_SUBTYPE_PKEY: + case MLX5_PORT_CHANGE_SUBTYPE_GUID: + case MLX5_PORT_CHANGE_SUBTYPE_CLIENT_REREG: + case MLX5_PORT_CHANGE_SUBTYPE_INITIALIZED: + dev_event = port_subtype2dev(eqe->sub_type); + dev_event_data = (unsigned long)port; + dev_event_dispatch = true; + break; + default: + mlx5_core_warn(dev, "Port event with unrecognized subtype: port %d, sub_type %d\n", + port, eqe->sub_type); + } + + if (dev->event && dev_event_dispatch) + dev->event(dev, dev_event, dev_event_data); + + return NOTIFY_OK; +} + +/* type == MLX5_EVENT_TYPE_GENERAL_EVENT */ +static int general_event(struct notifier_block *nb, unsigned long type, void *data) +{ + struct mlx5_event_nb *event_nb = mlx5_nb_cof(nb, struct mlx5_event_nb, nb); + struct mlx5_events *events = event_nb->ctx; + struct mlx5_core_dev *dev = events->dev; + + bool dev_event_dispatch = false; + enum mlx5_dev_event dev_event; + unsigned long dev_event_data; + struct mlx5_eqe *eqe = data; + + switch (eqe->sub_type) { + case MLX5_GENERAL_SUBTYPE_DELAY_DROP_TIMEOUT: + dev_event = MLX5_DEV_EVENT_DELAY_DROP_TIMEOUT; + dev_event_data = 0; + dev_event_dispatch = true; + break; + default: + mlx5_core_dbg(dev, "General event with unrecognized subtype: sub_type %d\n", + eqe->sub_type); + } + + if (dev->event && dev_event_dispatch) + dev->event(dev, dev_event, dev_event_data); + + return NOTIFY_OK; +} + +/* type == MLX5_EVENT_TYPE_TEMP_WARN_EVENT */ +static int temp_warn(struct notifier_block *nb, unsigned long type, void *data) +{ + struct mlx5_event_nb *event_nb = mlx5_nb_cof(nb, struct mlx5_event_nb, nb); + struct mlx5_events *events = event_nb->ctx; + struct mlx5_eqe *eqe = data; u64 value_lsb; u64 value_msb; value_lsb = be64_to_cpu(eqe->data.temp_warning.sensor_warning_lsb); value_msb = be64_to_cpu(eqe->data.temp_warning.sensor_warning_msb); - mlx5_core_warn(dev, + mlx5_core_warn(events->dev, "High temperature on sensors with bit set %llx %llx", value_msb, value_lsb); + + return NOTIFY_OK; } +/* MLX5_EVENT_TYPE_PORT_MODULE_EVENT */ static const char *mlx5_pme_status[MLX5_MODULE_STATUS_NUM] = { "Cable plugged", /* MLX5_MODULE_STATUS_PLUGGED = 0x1 */ "Cable unplugged", /* MLX5_MODULE_STATUS_UNPLUGGED = 0x2 */ @@ -132,12 +244,16 @@ static const char *mlx5_pme_error[MLX5_MODULE_EVENT_ERROR_NUM] = { "Unknown status", }; -static void port_module_event(struct mlx5_events *events, struct mlx5_eqe *eqe) +/* type == MLX5_EVENT_TYPE_PORT_MODULE_EVENT */ +static int port_module(struct notifier_block *nb, unsigned long type, void *data) { + struct mlx5_event_nb *event_nb = mlx5_nb_cof(nb, struct mlx5_event_nb, nb); + struct mlx5_events *events = event_nb->ctx; + struct mlx5_eqe *eqe = data; + enum port_module_event_status_type module_status; enum port_module_event_error_type error_type; struct mlx5_eqe_port_module *module_event_eqe; - struct mlx5_core_dev *dev = events->dev; u8 module_num; module_event_eqe = &eqe->data.port_module; @@ -146,7 +262,6 @@ static void port_module_event(struct mlx5_events *events, struct mlx5_eqe *eqe) PORT_MODULE_EVENT_MODULE_STATUS_MASK; error_type = module_event_eqe->error_type & PORT_MODULE_EVENT_ERROR_TYPE_MASK; - if (module_status < MLX5_MODULE_STATUS_ERROR) { events->pme_stats.status_counters[module_status - 1]++; } else if (module_status == MLX5_MODULE_STATUS_ERROR) { @@ -157,18 +272,20 @@ static void port_module_event(struct mlx5_events *events, struct mlx5_eqe *eqe) } if (!printk_ratelimit()) - return; + return NOTIFY_OK; if (module_status < MLX5_MODULE_STATUS_ERROR) - mlx5_core_info(dev, + mlx5_core_info(events->dev, "Port module event: module %u, %s\n", module_num, mlx5_pme_status[module_status - 1]); else if (module_status == MLX5_MODULE_STATUS_ERROR) - mlx5_core_info(dev, + mlx5_core_info(events->dev, "Port module event[error]: module %u, %s, %s\n", module_num, mlx5_pme_status[module_status - 1], mlx5_pme_error[error_type]); + + return NOTIFY_OK; } void mlx5_get_pme_stats(struct mlx5_core_dev *dev, struct mlx5_pme_stats *stats) @@ -176,80 +293,6 @@ void mlx5_get_pme_stats(struct mlx5_core_dev *dev, struct mlx5_pme_stats *stats) *stats = dev->priv.events->pme_stats; } -/* Event handler for the low level mlx5_core driver. - * This handler will process/filter _some_ events and sometimes dispatch - * the equivalent mlx5_dev_event to the HCA interfaces (mlx5_ib and mlx5e) - * - * Other Major feature specific events such as - * clock/eswitch/fpga/FW trace and many others, are handled elsewhere, with - * separate notifiers callbacks, specifically by those mlx5 components. - */ -static int events_notifier(struct notifier_block *nb, - unsigned long type, void *data) -{ - bool dev_event_dispatch = false; - enum mlx5_dev_event dev_event; - unsigned long dev_event_data; - - struct mlx5_eqe *eqe = data; - struct mlx5_events *events; - struct mlx5_core_dev *dev; - u8 port; - - events = mlx5_nb_cof(nb, struct mlx5_events, nb); - dev = events->dev; - - mlx5_core_dbg(dev, "Async eqe type %s, subtype (%d)\n", - eqe_type_str(eqe->type), eqe->sub_type); - switch (eqe->type) { - case MLX5_EVENT_TYPE_PORT_CHANGE: - port = (eqe->data.port.port >> 4) & 0xf; - switch (eqe->sub_type) { - case MLX5_PORT_CHANGE_SUBTYPE_DOWN: - case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE: - case MLX5_PORT_CHANGE_SUBTYPE_LID: - case MLX5_PORT_CHANGE_SUBTYPE_PKEY: - case MLX5_PORT_CHANGE_SUBTYPE_GUID: - case MLX5_PORT_CHANGE_SUBTYPE_CLIENT_REREG: - case MLX5_PORT_CHANGE_SUBTYPE_INITIALIZED: - dev_event = port_subtype2dev(eqe->sub_type); - dev_event_data = (unsigned long)port; - dev_event_dispatch = true; - break; - default: - mlx5_core_warn(dev, "Port event with unrecognized subtype: port %d, sub_type %d\n", - port, eqe->sub_type); - } - break; - case MLX5_EVENT_TYPE_GENERAL_EVENT: - switch (eqe->sub_type) { - case MLX5_GENERAL_SUBTYPE_DELAY_DROP_TIMEOUT: - dev_event = MLX5_DEV_EVENT_DELAY_DROP_TIMEOUT; - dev_event_data = 0; - dev_event_dispatch = true; - break; - default: - mlx5_core_dbg(dev, "General event with unrecognized subtype: sub_type %d\n", - eqe->sub_type); - } - break; - - case MLX5_EVENT_TYPE_PORT_MODULE_EVENT: - port_module_event(events, eqe); - break; - case MLX5_EVENT_TYPE_TEMP_WARN_EVENT: - temp_warning_event(dev, eqe); - break; - default: - return NOTIFY_DONE; - } - - if (dev->event && dev_event_dispatch) - dev->event(dev, dev_event, dev_event_data); - - return NOTIFY_OK; -} - int mlx5_events_init(struct mlx5_core_dev *dev) { struct mlx5_events *events = kzalloc(sizeof(*events), GFP_KERNEL); @@ -270,14 +313,20 @@ void mlx5_events_cleanup(struct mlx5_core_dev *dev) void mlx5_events_start(struct mlx5_core_dev *dev) { struct mlx5_events *events = dev->priv.events; + int i; - MLX5_NB_INIT(&events->nb, events_notifier, NOTIFY_ANY); - mlx5_eq_notifier_register(dev, &events->nb); + for (i = 0; i < ARRAY_SIZE(events_nbs_ref); i++) { + events->notifiers[i].nb = events_nbs_ref[i]; + events->notifiers[i].ctx = events; + mlx5_eq_notifier_register(dev, &events->notifiers[i].nb); + } } void mlx5_events_stop(struct mlx5_core_dev *dev) { struct mlx5_events *events = dev->priv.events; + int i; - mlx5_eq_notifier_unregister(dev, &events->nb); + for (i = ARRAY_SIZE(events_nbs_ref) - 1; i >= 0 ; i--) + mlx5_eq_notifier_unregister(dev, &events->notifiers[i].nb); } -- cgit v1.2.3-55-g7522 From 20902be46c4da59b1891d238801146134e0e06b5 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 26 Nov 2018 14:38:56 -0800 Subject: net/mlx5: Driver events notifier API Use atomic notifier chain to fire events to mlx5 core driver consumers (mlx5e/mlx5_ib) and provide mlx5 register/unregister notifier API. This API will replace the current mlx5_interface->event callback and all the logic around it, especially the delayed events logic introduced by commit 97834eba7c19 ("net/mlx5: Delay events till ib registration ends") Which is not needed anymore with this new API where the mlx5 interface can dynamically register/unregister its notifier. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/events.c | 25 +++++++++++++++++++++- drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h | 1 + include/linux/mlx5/driver.h | 4 ++++ 3 files changed, 29 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index 3ad004af37d7..560cc14c55f7 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -35,7 +35,8 @@ static struct mlx5_nb events_nbs_ref[] = { struct mlx5_events { struct mlx5_core_dev *dev; struct mlx5_event_nb notifiers[ARRAY_SIZE(events_nbs_ref)]; - + /* driver notifier chain */ + struct atomic_notifier_head nh; /* port module events stats */ struct mlx5_pme_stats pme_stats; }; @@ -300,6 +301,7 @@ int mlx5_events_init(struct mlx5_core_dev *dev) if (!events) return -ENOMEM; + ATOMIC_INIT_NOTIFIER_HEAD(&events->nh); events->dev = dev; dev->priv.events = events; return 0; @@ -330,3 +332,24 @@ void mlx5_events_stop(struct mlx5_core_dev *dev) for (i = ARRAY_SIZE(events_nbs_ref) - 1; i >= 0 ; i--) mlx5_eq_notifier_unregister(dev, &events->notifiers[i].nb); } + +int mlx5_notifier_register(struct mlx5_core_dev *dev, struct notifier_block *nb) +{ + struct mlx5_events *events = dev->priv.events; + + return atomic_notifier_chain_register(&events->nh, nb); +} +EXPORT_SYMBOL(mlx5_notifier_register); + +int mlx5_notifier_unregister(struct mlx5_core_dev *dev, struct notifier_block *nb) +{ + struct mlx5_events *events = dev->priv.events; + + return atomic_notifier_chain_unregister(&events->nh, nb); +} +EXPORT_SYMBOL(mlx5_notifier_unregister); + +int mlx5_notifier_call_chain(struct mlx5_events *events, unsigned int event, void *data) +{ + return atomic_notifier_call_chain(&events->nh, event, data); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h index 23317e328b0b..4d78a459676e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h @@ -73,5 +73,6 @@ struct mlx5_pme_stats { }; void mlx5_get_pme_stats(struct mlx5_core_dev *dev, struct mlx5_pme_stats *stats); +int mlx5_notifier_call_chain(struct mlx5_events *events, unsigned int event, void *data); #endif diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index ba64ecf72478..b96929d0cc9c 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -46,6 +46,7 @@ #include #include #include +#include #include #include @@ -1062,6 +1063,9 @@ struct mlx5_interface { void *mlx5_get_protocol_dev(struct mlx5_core_dev *mdev, int protocol); int mlx5_register_interface(struct mlx5_interface *intf); void mlx5_unregister_interface(struct mlx5_interface *intf); +int mlx5_notifier_register(struct mlx5_core_dev *dev, struct notifier_block *nb); +int mlx5_notifier_unregister(struct mlx5_core_dev *dev, struct notifier_block *nb); + int mlx5_core_query_vendor_id(struct mlx5_core_dev *mdev, u32 *vendor_id); int mlx5_cmd_create_vport_lag(struct mlx5_core_dev *dev); -- cgit v1.2.3-55-g7522 From 7a1795553074f90d59b3250b4a9b8aaf6ca2831e Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 26 Nov 2018 14:38:57 -0800 Subject: net/mlx5: Allow port change event to be forwarded to driver notifiers chain The idea is to allow mlx5 core interfaces (mlx5e/mlx5_ib) to be able to receive some allowed FW events as is via the new notifier API. In this patch we allow forwarding port change event to mlx5 core interfaces (mlx5e/mlx5_ib) as it was received from FW. Once mlx5e and mlx5_ib start using this event we can safely remove the redundant software version of it and its translation logic. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/events.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index 560cc14c55f7..adab66eb726c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -24,12 +24,18 @@ static int general_event(struct notifier_block *, unsigned long, void *); static int temp_warn(struct notifier_block *, unsigned long, void *); static int port_module(struct notifier_block *, unsigned long, void *); +/* handler which forwards the event to events->nh, driver notifiers */ +static int forward_event(struct notifier_block *, unsigned long, void *); + static struct mlx5_nb events_nbs_ref[] = { {.nb.notifier_call = any_notifier, .event_type = MLX5_EVENT_TYPE_NOTIFY_ANY }, {.nb.notifier_call = port_change, .event_type = MLX5_EVENT_TYPE_PORT_CHANGE }, {.nb.notifier_call = general_event, .event_type = MLX5_EVENT_TYPE_GENERAL_EVENT }, {.nb.notifier_call = temp_warn, .event_type = MLX5_EVENT_TYPE_TEMP_WARN_EVENT }, {.nb.notifier_call = port_module, .event_type = MLX5_EVENT_TYPE_PORT_MODULE_EVENT }, + + /* Events to be forwarded (as is) to mlx5 core interfaces (mlx5e/mlx5_ib) */ + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_PORT_CHANGE }, }; struct mlx5_events { @@ -294,6 +300,16 @@ void mlx5_get_pme_stats(struct mlx5_core_dev *dev, struct mlx5_pme_stats *stats) *stats = dev->priv.events->pme_stats; } +/* forward event as is to registered interfaces (mlx5e/mlx5_ib) */ +static int forward_event(struct notifier_block *nb, unsigned long event, void *data) +{ + struct mlx5_event_nb *event_nb = mlx5_nb_cof(nb, struct mlx5_event_nb, nb); + struct mlx5_events *events = event_nb->ctx; + + atomic_notifier_call_chain(&events->nh, event, data); + return NOTIFY_OK; +} + int mlx5_events_init(struct mlx5_core_dev *dev) { struct mlx5_events *events = kzalloc(sizeof(*events), GFP_KERNEL); -- cgit v1.2.3-55-g7522 From 7cffaddd39b438fab49863ea4602d6294a9d657f Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 26 Nov 2018 14:38:58 -0800 Subject: net/mlx5e: Use the new mlx5 core notifier API Remove the deprecated mlx5_interface->event mlx5e callback and use new mlx5 notifier API to subscribe for mlx5 events, handle port change event as received from FW rather than handling the mlx5 core processed port change software version event. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 3 ++- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 27 ++++++++++++----------- 2 files changed, 16 insertions(+), 14 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index aea74856c702..13d8a74d3db5 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -631,7 +631,6 @@ struct mlx5e_channel_stats { } ____cacheline_aligned_in_smp; enum { - MLX5E_STATE_ASYNC_EVENTS_ENABLED, MLX5E_STATE_OPENED, MLX5E_STATE_DESTROYING, }; @@ -690,6 +689,8 @@ struct mlx5e_priv { struct hwtstamp_config tstamp; u16 q_counter; u16 drop_rq_q_counter; + struct notifier_block events_nb; + #ifdef CONFIG_MLX5_CORE_EN_DCB struct mlx5e_dcbx dcbx; #endif diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 0d495a6b3949..56bc41b1c31f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -294,33 +294,35 @@ void mlx5e_queue_update_stats(struct mlx5e_priv *priv) queue_work(priv->wq, &priv->update_stats_work); } -static void mlx5e_async_event(struct mlx5_core_dev *mdev, void *vpriv, - enum mlx5_dev_event event, unsigned long param) +static int async_event(struct notifier_block *nb, unsigned long event, void *data) { - struct mlx5e_priv *priv = vpriv; + struct mlx5e_priv *priv = container_of(nb, struct mlx5e_priv, events_nb); + struct mlx5_eqe *eqe = data; - if (!test_bit(MLX5E_STATE_ASYNC_EVENTS_ENABLED, &priv->state)) - return; + if (event != MLX5_EVENT_TYPE_PORT_CHANGE) + return NOTIFY_DONE; - switch (event) { - case MLX5_DEV_EVENT_PORT_UP: - case MLX5_DEV_EVENT_PORT_DOWN: + switch (eqe->sub_type) { + case MLX5_PORT_CHANGE_SUBTYPE_DOWN: + case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE: queue_work(priv->wq, &priv->update_carrier_work); break; default: - break; + return NOTIFY_DONE; } + + return NOTIFY_OK; } static void mlx5e_enable_async_events(struct mlx5e_priv *priv) { - set_bit(MLX5E_STATE_ASYNC_EVENTS_ENABLED, &priv->state); + priv->events_nb.notifier_call = async_event; + mlx5_notifier_register(priv->mdev, &priv->events_nb); } static void mlx5e_disable_async_events(struct mlx5e_priv *priv) { - clear_bit(MLX5E_STATE_ASYNC_EVENTS_ENABLED, &priv->state); - mlx5_eq_synchronize_async_irq(priv->mdev); + mlx5_notifier_unregister(priv->mdev, &priv->events_nb); } static inline void mlx5e_build_umr_wqe(struct mlx5e_rq *rq, @@ -5170,7 +5172,6 @@ static struct mlx5_interface mlx5e_interface = { .remove = mlx5e_remove, .attach = mlx5e_attach, .detach = mlx5e_detach, - .event = mlx5e_async_event, .protocol = MLX5_INTERFACE_PROTOCOL_ETH, .get_dev = mlx5e_get_netdev, }; -- cgit v1.2.3-55-g7522 From 58d180b34e98698fec178a469b700f1bb5a32c1f Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 26 Nov 2018 14:38:59 -0800 Subject: net/mlx5: Forward all mlx5 events to mlx5 notifiers chain This to allow seamless migration to the new notifier chain API, and to eventually deprecate interfaces dev->event callback. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/dev.c | 3 +++ include/linux/mlx5/driver.h | 2 +- 2 files changed, 4 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c index 7eedbea38a78..d63ba8813829 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c @@ -32,6 +32,7 @@ #include #include "mlx5_core.h" +#include "lib/mlx5.h" static LIST_HEAD(intf_list); static LIST_HEAD(mlx5_dev_list); @@ -425,6 +426,8 @@ void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event, dev_ctx->intf->event(dev, dev_ctx->context, event, param); spin_unlock_irqrestore(&priv->ctx_lock, flags); + + mlx5_notifier_call_chain(dev->priv.events, event, (void *)param); } void mlx5_dev_list_lock(void) diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index b96929d0cc9c..14ca74707275 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -195,7 +195,7 @@ struct mlx5_rsc_debug { }; enum mlx5_dev_event { - MLX5_DEV_EVENT_SYS_ERROR, + MLX5_DEV_EVENT_SYS_ERROR = 128, /* 0 - 127 are FW events */ MLX5_DEV_EVENT_PORT_UP, MLX5_DEV_EVENT_PORT_DOWN, MLX5_DEV_EVENT_PORT_INITIALIZED, -- cgit v1.2.3-55-g7522 From 02039fb659b366011f55b15890136754f3d82e2d Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 26 Nov 2018 14:39:01 -0800 Subject: net/mlx5: Remove unused events callback and logic The mlx5_interface->event callback is not used by mlx5e/mlx5_ib anymore. We totally remove the delayed events logic work around, since with the dynamic notifier registration API it is not needed anymore, mlx5_ib can register its notifier and start receiving events exactly at the moment it is ready to handle them. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/dev.c | 109 +-------------------- drivers/net/ethernet/mellanox/mlx5/core/events.c | 8 +- drivers/net/ethernet/mellanox/mlx5/core/health.c | 3 +- drivers/net/ethernet/mellanox/mlx5/core/main.c | 10 -- .../net/ethernet/mellanox/mlx5/core/mlx5_core.h | 3 - include/linux/mlx5/driver.h | 10 +- 6 files changed, 11 insertions(+), 132 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c index d63ba8813829..d2ed14bc37c3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c @@ -32,7 +32,6 @@ #include #include "mlx5_core.h" -#include "lib/mlx5.h" static LIST_HEAD(intf_list); static LIST_HEAD(mlx5_dev_list); @@ -46,75 +45,11 @@ struct mlx5_device_context { unsigned long state; }; -struct mlx5_delayed_event { - struct list_head list; - struct mlx5_core_dev *dev; - enum mlx5_dev_event event; - unsigned long param; -}; - enum { MLX5_INTERFACE_ADDED, MLX5_INTERFACE_ATTACHED, }; -static void add_delayed_event(struct mlx5_priv *priv, - struct mlx5_core_dev *dev, - enum mlx5_dev_event event, - unsigned long param) -{ - struct mlx5_delayed_event *delayed_event; - - delayed_event = kzalloc(sizeof(*delayed_event), GFP_ATOMIC); - if (!delayed_event) { - mlx5_core_err(dev, "event %d is missed\n", event); - return; - } - - mlx5_core_dbg(dev, "Accumulating event %d\n", event); - delayed_event->dev = dev; - delayed_event->event = event; - delayed_event->param = param; - list_add_tail(&delayed_event->list, &priv->waiting_events_list); -} - -static void delayed_event_release(struct mlx5_device_context *dev_ctx, - struct mlx5_priv *priv) -{ - struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv); - struct mlx5_delayed_event *de; - struct mlx5_delayed_event *n; - struct list_head temp; - - INIT_LIST_HEAD(&temp); - - spin_lock_irq(&priv->ctx_lock); - - priv->is_accum_events = false; - list_splice_init(&priv->waiting_events_list, &temp); - if (!dev_ctx->context) - goto out; - list_for_each_entry_safe(de, n, &temp, list) - dev_ctx->intf->event(dev, dev_ctx->context, de->event, de->param); - -out: - spin_unlock_irq(&priv->ctx_lock); - - list_for_each_entry_safe(de, n, &temp, list) { - list_del(&de->list); - kfree(de); - } -} - -/* accumulating events that can come after mlx5_ib calls to - * ib_register_device, till adding that interface to the events list. - */ -static void delayed_event_start(struct mlx5_priv *priv) -{ - spin_lock_irq(&priv->ctx_lock); - priv->is_accum_events = true; - spin_unlock_irq(&priv->ctx_lock); -} void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv) { @@ -130,8 +65,6 @@ void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv) dev_ctx->intf = intf; - delayed_event_start(priv); - dev_ctx->context = intf->add(dev); if (dev_ctx->context) { set_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state); @@ -143,8 +76,6 @@ void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv) spin_unlock_irq(&priv->ctx_lock); } - delayed_event_release(dev_ctx, priv); - if (!dev_ctx->context) kfree(dev_ctx); } @@ -188,26 +119,20 @@ static void mlx5_attach_interface(struct mlx5_interface *intf, struct mlx5_priv if (!dev_ctx) return; - delayed_event_start(priv); if (intf->attach) { if (test_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state)) - goto out; + return; if (intf->attach(dev, dev_ctx->context)) - goto out; - + return; set_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state); } else { if (test_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state)) - goto out; + return; dev_ctx->context = intf->add(dev); if (!dev_ctx->context) - goto out; - + return; set_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state); } - -out: - delayed_event_release(dev_ctx, priv); } void mlx5_attach_device(struct mlx5_core_dev *dev) @@ -403,32 +328,6 @@ struct mlx5_core_dev *mlx5_get_next_phys_dev(struct mlx5_core_dev *dev) return res; } -void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event, - unsigned long param) -{ - struct mlx5_priv *priv = &dev->priv; - struct mlx5_device_context *dev_ctx; - unsigned long flags; - - spin_lock_irqsave(&priv->ctx_lock, flags); - - if (priv->is_accum_events) - add_delayed_event(priv, dev, event, param); - - /* After mlx5_detach_device, the dev_ctx->intf is still set and dev_ctx is - * still in priv->ctx_list. In this case, only notify the dev_ctx if its - * ADDED or ATTACHED bit are set. - */ - list_for_each_entry(dev_ctx, &priv->ctx_list, list) - if (dev_ctx->intf->event && - (test_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state) || - test_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state))) - dev_ctx->intf->event(dev, dev_ctx->context, event, param); - - spin_unlock_irqrestore(&priv->ctx_lock, flags); - - mlx5_notifier_call_chain(dev->priv.events, event, (void *)param); -} void mlx5_dev_list_lock(void) { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index adab66eb726c..ab66f5d65a04 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -178,8 +178,8 @@ static int port_change(struct notifier_block *nb, port, eqe->sub_type); } - if (dev->event && dev_event_dispatch) - dev->event(dev, dev_event, dev_event_data); + if (dev_event_dispatch) + mlx5_notifier_call_chain(events, dev_event, (void *)dev_event_data); return NOTIFY_OK; } @@ -207,8 +207,8 @@ static int general_event(struct notifier_block *nb, unsigned long type, void *da eqe->sub_type); } - if (dev->event && dev_event_dispatch) - dev->event(dev, dev_event, dev_event_data); + if (dev_event_dispatch) + mlx5_notifier_call_chain(events, dev_event, (void *)dev_event_data); return NOTIFY_OK; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c index 4e42bd290959..196c07383082 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/health.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c @@ -39,6 +39,7 @@ #include #include "mlx5_core.h" #include "lib/eq.h" +#include "lib/mlx5.h" enum { MLX5_HEALTH_POLL_INTERVAL = 2 * HZ, @@ -105,7 +106,7 @@ void mlx5_enter_error_state(struct mlx5_core_dev *dev, bool force) mlx5_cmd_trigger_completions(dev); } - mlx5_core_event(dev, MLX5_DEV_EVENT_SYS_ERROR, 1); + mlx5_notifier_call_chain(dev->priv.events, MLX5_DEV_EVENT_SYS_ERROR, (void *)1); mlx5_core_err(dev, "end\n"); unlock: diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index e56278ead4eb..4bc27a073dc4 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -1125,12 +1125,6 @@ out: return err; } -struct mlx5_core_event_handler { - void (*event)(struct mlx5_core_dev *dev, - enum mlx5_dev_event event, - void *data); -}; - static const struct devlink_ops mlx5_devlink_ops = { #ifdef CONFIG_MLX5_ESWITCH .eswitch_mode_set = mlx5_devlink_eswitch_mode_set, @@ -1164,7 +1158,6 @@ static int init_one(struct pci_dev *pdev, pci_set_drvdata(pdev, dev); dev->pdev = pdev; - dev->event = mlx5_core_event; dev->profile = &profile[prof_sel]; INIT_LIST_HEAD(&priv->ctx_list); @@ -1172,9 +1165,6 @@ static int init_one(struct pci_dev *pdev, mutex_init(&dev->pci_status_mutex); mutex_init(&dev->intf_state_mutex); - INIT_LIST_HEAD(&priv->waiting_events_list); - priv->is_accum_events = false; - mutex_init(&priv->bfregs.reg_head.lock); mutex_init(&priv->bfregs.wc_head.lock); INIT_LIST_HEAD(&priv->bfregs.reg_head.list); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index c70bd94e18d6..fd3141a4b3f1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -102,9 +102,6 @@ int mlx5_cmd_init_hca(struct mlx5_core_dev *dev, uint32_t *sw_owner_id); int mlx5_cmd_teardown_hca(struct mlx5_core_dev *dev); int mlx5_cmd_force_teardown_hca(struct mlx5_core_dev *dev); int mlx5_cmd_fast_teardown_hca(struct mlx5_core_dev *dev); - -void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event, - unsigned long param); void mlx5_enter_error_state(struct mlx5_core_dev *dev, bool force); void mlx5_disable_device(struct mlx5_core_dev *dev); void mlx5_recover_device(struct mlx5_core_dev *dev); diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 14ca74707275..d3ffc64f9a75 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -588,10 +588,7 @@ struct mlx5_priv { struct list_head dev_list; struct list_head ctx_list; spinlock_t ctx_lock; - - struct list_head waiting_events_list; - bool is_accum_events; - struct mlx5_events *events; + struct mlx5_events *events; struct mlx5_flow_steering *steering; struct mlx5_mpfs *mpfs; @@ -696,9 +693,6 @@ struct mlx5_core_dev { /* sync interface state */ struct mutex intf_state_mutex; unsigned long intf_state; - void (*event) (struct mlx5_core_dev *dev, - enum mlx5_dev_event event, - unsigned long param); struct mlx5_priv priv; struct mlx5_profile *profile; atomic_t num_qps; @@ -1053,8 +1047,6 @@ struct mlx5_interface { void (*remove)(struct mlx5_core_dev *dev, void *context); int (*attach)(struct mlx5_core_dev *dev, void *context); void (*detach)(struct mlx5_core_dev *dev, void *context); - void (*event)(struct mlx5_core_dev *dev, void *context, - enum mlx5_dev_event event, unsigned long param); void * (*get_dev)(void *context); int protocol; struct list_head list; -- cgit v1.2.3-55-g7522 From cb6191bf25f755f47593fc40b8fd39d3eedde884 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 26 Nov 2018 14:39:03 -0800 Subject: net/mlx5: Allow forwarding event type general event as is FW general event is used by mlx5_ib for RQ delay drop timeout event handling, in this patch we allow to forward FW general event type to mlx5 notifiers chain so mlx5_ib can handle it and to deprecate the software version of it. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/events.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index ab66f5d65a04..735a9b038a73 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -36,6 +36,7 @@ static struct mlx5_nb events_nbs_ref[] = { /* Events to be forwarded (as is) to mlx5 core interfaces (mlx5e/mlx5_ib) */ {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_PORT_CHANGE }, + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_GENERAL_EVENT }, }; struct mlx5_events { -- cgit v1.2.3-55-g7522 From b8267cd765b333673e05696b517d38a1a7eb5b2e Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 26 Nov 2018 14:39:05 -0800 Subject: net/mlx5: Remove all deprecated software versions of FW events Before the new mlx5 event notification infrastructure and API, mlx5_core used to process all events before forwarding them to mlx5 interfaces (mlx5e/mlx5_ib) and used to translate the event type enum to a software defined enum, this is not needed anymore since it is ok for mlx5e and mlx5_ib to receive FW events as is, at least the few ones mlx5 core allows. mlx5e and mlx5_ib already moved to use the new API and they only handle FW events types, it is now safe to remove all equivalent software defined events and the logic around them. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/events.c | 92 +----------------------- include/linux/mlx5/driver.h | 9 --- 2 files changed, 1 insertion(+), 100 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index 735a9b038a73..3708b42c1d6b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -19,8 +19,6 @@ struct mlx5_event_nb { * separate notifiers callbacks, specifically by those mlx5 components. */ static int any_notifier(struct notifier_block *, unsigned long, void *); -static int port_change(struct notifier_block *, unsigned long, void *); -static int general_event(struct notifier_block *, unsigned long, void *); static int temp_warn(struct notifier_block *, unsigned long, void *); static int port_module(struct notifier_block *, unsigned long, void *); @@ -28,9 +26,8 @@ static int port_module(struct notifier_block *, unsigned long, void *); static int forward_event(struct notifier_block *, unsigned long, void *); static struct mlx5_nb events_nbs_ref[] = { + /* Events to be proccessed by mlx5_core */ {.nb.notifier_call = any_notifier, .event_type = MLX5_EVENT_TYPE_NOTIFY_ANY }, - {.nb.notifier_call = port_change, .event_type = MLX5_EVENT_TYPE_PORT_CHANGE }, - {.nb.notifier_call = general_event, .event_type = MLX5_EVENT_TYPE_GENERAL_EVENT }, {.nb.notifier_call = temp_warn, .event_type = MLX5_EVENT_TYPE_TEMP_WARN_EVENT }, {.nb.notifier_call = port_module, .event_type = MLX5_EVENT_TYPE_PORT_MODULE_EVENT }, @@ -127,93 +124,6 @@ static int any_notifier(struct notifier_block *nb, return NOTIFY_OK; } -static enum mlx5_dev_event port_subtype2dev(u8 subtype) -{ - switch (subtype) { - case MLX5_PORT_CHANGE_SUBTYPE_DOWN: - return MLX5_DEV_EVENT_PORT_DOWN; - case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE: - return MLX5_DEV_EVENT_PORT_UP; - case MLX5_PORT_CHANGE_SUBTYPE_INITIALIZED: - return MLX5_DEV_EVENT_PORT_INITIALIZED; - case MLX5_PORT_CHANGE_SUBTYPE_LID: - return MLX5_DEV_EVENT_LID_CHANGE; - case MLX5_PORT_CHANGE_SUBTYPE_PKEY: - return MLX5_DEV_EVENT_PKEY_CHANGE; - case MLX5_PORT_CHANGE_SUBTYPE_GUID: - return MLX5_DEV_EVENT_GUID_CHANGE; - case MLX5_PORT_CHANGE_SUBTYPE_CLIENT_REREG: - return MLX5_DEV_EVENT_CLIENT_REREG; - } - return -1; -} - -/* type == MLX5_EVENT_TYPE_PORT_CHANGE */ -static int port_change(struct notifier_block *nb, - unsigned long type, void *data) -{ - struct mlx5_event_nb *event_nb = mlx5_nb_cof(nb, struct mlx5_event_nb, nb); - struct mlx5_events *events = event_nb->ctx; - struct mlx5_core_dev *dev = events->dev; - - bool dev_event_dispatch = false; - enum mlx5_dev_event dev_event; - unsigned long dev_event_data; - struct mlx5_eqe *eqe = data; - u8 port = (eqe->data.port.port >> 4) & 0xf; - - switch (eqe->sub_type) { - case MLX5_PORT_CHANGE_SUBTYPE_DOWN: - case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE: - case MLX5_PORT_CHANGE_SUBTYPE_LID: - case MLX5_PORT_CHANGE_SUBTYPE_PKEY: - case MLX5_PORT_CHANGE_SUBTYPE_GUID: - case MLX5_PORT_CHANGE_SUBTYPE_CLIENT_REREG: - case MLX5_PORT_CHANGE_SUBTYPE_INITIALIZED: - dev_event = port_subtype2dev(eqe->sub_type); - dev_event_data = (unsigned long)port; - dev_event_dispatch = true; - break; - default: - mlx5_core_warn(dev, "Port event with unrecognized subtype: port %d, sub_type %d\n", - port, eqe->sub_type); - } - - if (dev_event_dispatch) - mlx5_notifier_call_chain(events, dev_event, (void *)dev_event_data); - - return NOTIFY_OK; -} - -/* type == MLX5_EVENT_TYPE_GENERAL_EVENT */ -static int general_event(struct notifier_block *nb, unsigned long type, void *data) -{ - struct mlx5_event_nb *event_nb = mlx5_nb_cof(nb, struct mlx5_event_nb, nb); - struct mlx5_events *events = event_nb->ctx; - struct mlx5_core_dev *dev = events->dev; - - bool dev_event_dispatch = false; - enum mlx5_dev_event dev_event; - unsigned long dev_event_data; - struct mlx5_eqe *eqe = data; - - switch (eqe->sub_type) { - case MLX5_GENERAL_SUBTYPE_DELAY_DROP_TIMEOUT: - dev_event = MLX5_DEV_EVENT_DELAY_DROP_TIMEOUT; - dev_event_data = 0; - dev_event_dispatch = true; - break; - default: - mlx5_core_dbg(dev, "General event with unrecognized subtype: sub_type %d\n", - eqe->sub_type); - } - - if (dev_event_dispatch) - mlx5_notifier_call_chain(events, dev_event, (void *)dev_event_data); - - return NOTIFY_OK; -} - /* type == MLX5_EVENT_TYPE_TEMP_WARN_EVENT */ static int temp_warn(struct notifier_block *nb, unsigned long type, void *data) { diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index d3ffc64f9a75..a77bedb8a556 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -196,15 +196,6 @@ struct mlx5_rsc_debug { enum mlx5_dev_event { MLX5_DEV_EVENT_SYS_ERROR = 128, /* 0 - 127 are FW events */ - MLX5_DEV_EVENT_PORT_UP, - MLX5_DEV_EVENT_PORT_DOWN, - MLX5_DEV_EVENT_PORT_INITIALIZED, - MLX5_DEV_EVENT_LID_CHANGE, - MLX5_DEV_EVENT_PKEY_CHANGE, - MLX5_DEV_EVENT_GUID_CHANGE, - MLX5_DEV_EVENT_CLIENT_REREG, - MLX5_DEV_EVENT_PPS, - MLX5_DEV_EVENT_DELAY_DROP_TIMEOUT, }; enum mlx5_port_status { -- cgit v1.2.3-55-g7522 From 451be51c0b474f790e9833cd575fd9a6fbd679df Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 26 Nov 2018 14:39:06 -0800 Subject: net/mlx5: Forward QP/WorkQueues resource events Allow forwarding QP and WQ events to mlx5_core interfaces, e.g. mlx5_ib Use mlx5_notifier_register/unregister in qp.c in order to allow seamless transition of qp.c to infiniband subsystem. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/events.c | 10 ++++++++++ drivers/net/ethernet/mellanox/mlx5/core/qp.c | 8 ++++---- include/linux/mlx5/driver.h | 2 +- 3 files changed, 15 insertions(+), 5 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index 3708b42c1d6b..201c5f6091ea 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -34,6 +34,16 @@ static struct mlx5_nb events_nbs_ref[] = { /* Events to be forwarded (as is) to mlx5 core interfaces (mlx5e/mlx5_ib) */ {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_PORT_CHANGE }, {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_GENERAL_EVENT }, + /* QP/WQ resource events to forward */ + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_DCT_DRAINED }, + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_PATH_MIG }, + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_COMM_EST }, + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_SQ_DRAINED }, + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_SRQ_LAST_WQE }, + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_WQ_CATAS_ERROR }, + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_PATH_MIG_FAILED }, + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR }, + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_WQ_ACCESS_ERROR }, }; struct mlx5_events { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/qp.c b/drivers/net/ethernet/mellanox/mlx5/core/qp.c index 28726c63101f..388f205a497f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/qp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/qp.c @@ -150,7 +150,7 @@ static int rsc_event_notifier(struct notifier_block *nb, return NOTIFY_DONE; } - table = mlx5_nb_cof(nb, struct mlx5_qp_table, nb); + table = container_of(nb, struct mlx5_qp_table, nb); priv = container_of(table, struct mlx5_priv, qp_table); dev = container_of(priv, struct mlx5_core_dev, priv); @@ -523,15 +523,15 @@ void mlx5_init_qp_table(struct mlx5_core_dev *dev) INIT_RADIX_TREE(&table->tree, GFP_ATOMIC); mlx5_qp_debugfs_init(dev); - MLX5_NB_INIT(&table->nb, rsc_event_notifier, NOTIFY_ANY); - mlx5_eq_notifier_register(dev, &table->nb); + table->nb.notifier_call = rsc_event_notifier; + mlx5_notifier_register(dev, &table->nb); } void mlx5_cleanup_qp_table(struct mlx5_core_dev *dev) { struct mlx5_qp_table *table = &dev->priv.qp_table; - mlx5_eq_notifier_unregister(dev, &table->nb); + mlx5_notifier_unregister(dev, &table->nb); mlx5_qp_debugfs_cleanup(dev); } diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index a77bedb8a556..4f078b7f6620 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -456,7 +456,7 @@ struct mlx5_core_health { }; struct mlx5_qp_table { - struct mlx5_nb nb; + struct notifier_block nb; /* protect radix tree */ -- cgit v1.2.3-55-g7522 From 4e2df04ad25ab8e627878817e56d6a27645ca4a8 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 26 Nov 2018 14:39:07 -0800 Subject: net/mlx5: Forward SRQ resource events Allow forwarding of SRQ events to mlx5_core interfaces, e.g. mlx5_ib. Use mlx5_notifier_register/unregister in srq.c in order to allow seamless transition of srq.c to infiniband subsystem. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/events.c | 3 ++ drivers/net/ethernet/mellanox/mlx5/core/srq.c | 38 +++++++----------------- include/linux/mlx5/driver.h | 3 +- 3 files changed, 14 insertions(+), 30 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index 201c5f6091ea..9e6e216faac3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -44,6 +44,9 @@ static struct mlx5_nb events_nbs_ref[] = { {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_PATH_MIG_FAILED }, {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR }, {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_WQ_ACCESS_ERROR }, + /* SRQ events */ + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_SRQ_CATAS_ERROR }, + {.nb.notifier_call = forward_event, .event_type = MLX5_EVENT_TYPE_SRQ_RQ_LIMIT }, }; struct mlx5_events { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/srq.c b/drivers/net/ethernet/mellanox/mlx5/core/srq.c index 0563866c13f2..79c5f0d57956 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/srq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/srq.c @@ -40,15 +40,21 @@ #include "mlx5_core.h" #include "lib/eq.h" -static int srq_event_notifier(struct mlx5_srq_table *table, +static int srq_event_notifier(struct notifier_block *nb, unsigned long type, void *data) { + struct mlx5_srq_table *table; struct mlx5_core_dev *dev; struct mlx5_core_srq *srq; struct mlx5_priv *priv; struct mlx5_eqe *eqe; u32 srqn; + if (type != MLX5_EVENT_TYPE_SRQ_CATAS_ERROR && + type != MLX5_EVENT_TYPE_SRQ_RQ_LIMIT) + return NOTIFY_DONE; + + table = container_of(nb, struct mlx5_srq_table, nb); priv = container_of(table, struct mlx5_priv, srq_table); dev = container_of(priv, struct mlx5_core_dev, priv); @@ -77,26 +83,6 @@ static int srq_event_notifier(struct mlx5_srq_table *table, return NOTIFY_OK; } -static int catas_err_notifier(struct notifier_block *nb, - unsigned long type, void *data) -{ - struct mlx5_srq_table *table; - - table = mlx5_nb_cof(nb, struct mlx5_srq_table, catas_err_nb); - /* type == MLX5_EVENT_TYPE_SRQ_CATAS_ERROR */ - return srq_event_notifier(table, type, data); -} - -static int rq_limit_notifier(struct notifier_block *nb, - unsigned long type, void *data) -{ - struct mlx5_srq_table *table; - - table = mlx5_nb_cof(nb, struct mlx5_srq_table, rq_limit_nb); - /* type == MLX5_EVENT_TYPE_SRQ_RQ_LIMIT */ - return srq_event_notifier(table, type, data); -} - static int get_pas_size(struct mlx5_srq_attr *in) { u32 log_page_size = in->log_page_size + 12; @@ -743,17 +729,13 @@ void mlx5_init_srq_table(struct mlx5_core_dev *dev) spin_lock_init(&table->lock); INIT_RADIX_TREE(&table->tree, GFP_ATOMIC); - MLX5_NB_INIT(&table->catas_err_nb, catas_err_notifier, SRQ_CATAS_ERROR); - mlx5_eq_notifier_register(dev, &table->catas_err_nb); - - MLX5_NB_INIT(&table->rq_limit_nb, rq_limit_notifier, SRQ_RQ_LIMIT); - mlx5_eq_notifier_register(dev, &table->rq_limit_nb); + table->nb.notifier_call = srq_event_notifier; + mlx5_notifier_register(dev, &table->nb); } void mlx5_cleanup_srq_table(struct mlx5_core_dev *dev) { struct mlx5_srq_table *table = &dev->priv.srq_table; - mlx5_eq_notifier_unregister(dev, &table->rq_limit_nb); - mlx5_eq_notifier_unregister(dev, &table->catas_err_nb); + mlx5_notifier_unregister(dev, &table->nb); } diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 4f078b7f6620..27a481b159ed 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -465,8 +465,7 @@ struct mlx5_qp_table { }; struct mlx5_srq_table { - struct mlx5_nb catas_err_nb; - struct mlx5_nb rq_limit_nb; + struct notifier_block nb; /* protect radix tree */ spinlock_t lock; -- cgit v1.2.3-55-g7522 From 93631211c9c07dd3e60ca8462281ff475d3b87a6 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 26 Nov 2018 14:39:08 -0800 Subject: net/mlx5: Debug print for forwarded async events Print a debug message for every async FW event forwarded to mlx5 interfaces (mlx5e netdev and mlx5_ib rdma module). Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/events.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index 9e6e216faac3..e92df7020a26 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -229,7 +229,10 @@ static int forward_event(struct notifier_block *nb, unsigned long event, void *d { struct mlx5_event_nb *event_nb = mlx5_nb_cof(nb, struct mlx5_event_nb, nb); struct mlx5_events *events = event_nb->ctx; + struct mlx5_eqe *eqe = data; + mlx5_core_dbg(events->dev, "Async eqe type %s, subtype (%d) forward to interfaces\n", + eqe_type_str(eqe->type), eqe->sub_type); atomic_notifier_call_chain(&events->nh, event, data); return NOTIFY_OK; } -- cgit v1.2.3-55-g7522 From 6cd0014ab90f6959fa1f8cc8b3f38d302457c919 Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Wed, 28 Nov 2018 20:53:33 +0200 Subject: net/mlx5: Align SRQ licenses and copyright information Ensure that both RDMA and netdev parts of SRQ implementation has same copyright and license information annotated by SPDX tags. Reviewed-by: Mark Bloch Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/mlx5/srq.c | 31 ++------------------------- drivers/net/ethernet/mellanox/mlx5/core/srq.c | 31 ++------------------------- include/linux/mlx5/srq.h | 31 ++------------------------- 3 files changed, 6 insertions(+), 87 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/infiniband/hw/mlx5/srq.c b/drivers/infiniband/hw/mlx5/srq.c index d012e7dbcc38..28794780062e 100644 --- a/drivers/infiniband/hw/mlx5/srq.c +++ b/drivers/infiniband/hw/mlx5/srq.c @@ -1,33 +1,6 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB /* - * Copyright (c) 2013-2015, Mellanox Technologies. All rights reserved. - * - * This software is available to you under a choice of one of two - * licenses. You may choose to be licensed under the terms of the GNU - * General Public License (GPL) Version 2, available from the file - * COPYING in the main directory of this source tree, or the - * OpenIB.org BSD license below: - * - * Redistribution and use in source and binary forms, with or - * without modification, are permitted provided that the following - * conditions are met: - * - * - Redistributions of source code must retain the above - * copyright notice, this list of conditions and the following - * disclaimer. - * - * - Redistributions in binary form must reproduce the above - * copyright notice, this list of conditions and the following - * disclaimer in the documentation and/or other materials - * provided with the distribution. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, - * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND - * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS - * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN - * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE - * SOFTWARE. + * Copyright (c) 2013-2018, Mellanox Technologies inc. All rights reserved. */ #include diff --git a/drivers/net/ethernet/mellanox/mlx5/core/srq.c b/drivers/net/ethernet/mellanox/mlx5/core/srq.c index 79c5f0d57956..10036aaa200a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/srq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/srq.c @@ -1,33 +1,6 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB /* - * Copyright (c) 2013-2015, Mellanox Technologies. All rights reserved. - * - * This software is available to you under a choice of one of two - * licenses. You may choose to be licensed under the terms of the GNU - * General Public License (GPL) Version 2, available from the file - * COPYING in the main directory of this source tree, or the - * OpenIB.org BSD license below: - * - * Redistribution and use in source and binary forms, with or - * without modification, are permitted provided that the following - * conditions are met: - * - * - Redistributions of source code must retain the above - * copyright notice, this list of conditions and the following - * disclaimer. - * - * - Redistributions in binary form must reproduce the above - * copyright notice, this list of conditions and the following - * disclaimer in the documentation and/or other materials - * provided with the distribution. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, - * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND - * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS - * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN - * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE - * SOFTWARE. + * Copyright (c) 2013-2018, Mellanox Technologies inc. All rights reserved. */ #include diff --git a/include/linux/mlx5/srq.h b/include/linux/mlx5/srq.h index 1b1f3c20c6a3..77bc4264066d 100644 --- a/include/linux/mlx5/srq.h +++ b/include/linux/mlx5/srq.h @@ -1,33 +1,6 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ /* - * Copyright (c) 2013-2015, Mellanox Technologies. All rights reserved. - * - * This software is available to you under a choice of one of two - * licenses. You may choose to be licensed under the terms of the GNU - * General Public License (GPL) Version 2, available from the file - * COPYING in the main directory of this source tree, or the - * OpenIB.org BSD license below: - * - * Redistribution and use in source and binary forms, with or - * without modification, are permitted provided that the following - * conditions are met: - * - * - Redistributions of source code must retain the above - * copyright notice, this list of conditions and the following - * disclaimer. - * - * - Redistributions in binary form must reproduce the above - * copyright notice, this list of conditions and the following - * disclaimer in the documentation and/or other materials - * provided with the distribution. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, - * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND - * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS - * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN - * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE - * SOFTWARE. + * Copyright (c) 2013-2018, Mellanox Technologies. All rights reserved. */ #ifndef MLX5_SRQ_H -- cgit v1.2.3-55-g7522 From 5b5f0f16276021794038f12adc56df70cec42b4f Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Wed, 28 Nov 2018 20:53:34 +0200 Subject: net/mlx5: Remove dead transobj code Delete functions which are not called and not needed. Reviewed-by: Mark Bloch Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/transobj.c | 66 ---------------------- include/linux/mlx5/transobj.h | 5 -- 2 files changed, 71 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/transobj.c b/drivers/net/ethernet/mellanox/mlx5/core/transobj.c index a1ee9a8a769e..ab482124e901 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/transobj.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/transobj.c @@ -301,72 +301,6 @@ int mlx5_core_query_rmp(struct mlx5_core_dev *dev, u32 rmpn, u32 *out) return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen); } -int mlx5_core_arm_rmp(struct mlx5_core_dev *dev, u32 rmpn, u16 lwm) -{ - void *in; - void *rmpc; - void *wq; - void *bitmask; - int err; - - in = kvzalloc(MLX5_ST_SZ_BYTES(modify_rmp_in), GFP_KERNEL); - if (!in) - return -ENOMEM; - - rmpc = MLX5_ADDR_OF(modify_rmp_in, in, ctx); - bitmask = MLX5_ADDR_OF(modify_rmp_in, in, bitmask); - wq = MLX5_ADDR_OF(rmpc, rmpc, wq); - - MLX5_SET(modify_rmp_in, in, rmp_state, MLX5_RMPC_STATE_RDY); - MLX5_SET(modify_rmp_in, in, rmpn, rmpn); - MLX5_SET(wq, wq, lwm, lwm); - MLX5_SET(rmp_bitmask, bitmask, lwm, 1); - MLX5_SET(rmpc, rmpc, state, MLX5_RMPC_STATE_RDY); - - err = mlx5_core_modify_rmp(dev, in, MLX5_ST_SZ_BYTES(modify_rmp_in)); - - kvfree(in); - - return err; -} - -int mlx5_core_create_xsrq(struct mlx5_core_dev *dev, u32 *in, int inlen, - u32 *xsrqn) -{ - u32 out[MLX5_ST_SZ_DW(create_xrc_srq_out)] = {0}; - int err; - - MLX5_SET(create_xrc_srq_in, in, opcode, MLX5_CMD_OP_CREATE_XRC_SRQ); - err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out)); - if (!err) - *xsrqn = MLX5_GET(create_xrc_srq_out, out, xrc_srqn); - - return err; -} - -int mlx5_core_destroy_xsrq(struct mlx5_core_dev *dev, u32 xsrqn) -{ - u32 in[MLX5_ST_SZ_DW(destroy_xrc_srq_in)] = {0}; - u32 out[MLX5_ST_SZ_DW(destroy_xrc_srq_out)] = {0}; - - MLX5_SET(destroy_xrc_srq_in, in, opcode, MLX5_CMD_OP_DESTROY_XRC_SRQ); - MLX5_SET(destroy_xrc_srq_in, in, xrc_srqn, xsrqn); - return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); -} - -int mlx5_core_arm_xsrq(struct mlx5_core_dev *dev, u32 xsrqn, u16 lwm) -{ - u32 in[MLX5_ST_SZ_DW(arm_xrc_srq_in)] = {0}; - u32 out[MLX5_ST_SZ_DW(arm_xrc_srq_out)] = {0}; - - MLX5_SET(arm_xrc_srq_in, in, opcode, MLX5_CMD_OP_ARM_XRC_SRQ); - MLX5_SET(arm_xrc_srq_in, in, xrc_srqn, xsrqn); - MLX5_SET(arm_xrc_srq_in, in, lwm, lwm); - MLX5_SET(arm_xrc_srq_in, in, op_mod, - MLX5_ARM_XRC_SRQ_IN_OP_MOD_XRC_SRQ); - return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); -} - int mlx5_core_create_rqt(struct mlx5_core_dev *dev, u32 *in, int inlen, u32 *rqtn) { diff --git a/include/linux/mlx5/transobj.h b/include/linux/mlx5/transobj.h index 7f5ca2cd3a32..39ebb699875b 100644 --- a/include/linux/mlx5/transobj.h +++ b/include/linux/mlx5/transobj.h @@ -63,11 +63,6 @@ int mlx5_core_create_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen, int mlx5_core_modify_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen); int mlx5_core_destroy_rmp(struct mlx5_core_dev *dev, u32 rmpn); int mlx5_core_query_rmp(struct mlx5_core_dev *dev, u32 rmpn, u32 *out); -int mlx5_core_arm_rmp(struct mlx5_core_dev *dev, u32 rmpn, u16 lwm); -int mlx5_core_create_xsrq(struct mlx5_core_dev *dev, u32 *in, int inlen, - u32 *rmpn); -int mlx5_core_destroy_xsrq(struct mlx5_core_dev *dev, u32 rmpn); -int mlx5_core_arm_xsrq(struct mlx5_core_dev *dev, u32 rmpn, u16 lwm); int mlx5_core_create_rqt(struct mlx5_core_dev *dev, u32 *in, int inlen, u32 *rqtn); -- cgit v1.2.3-55-g7522 From 26d1164dff81096c38d955f00b84a0075563d8c7 Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Wed, 28 Nov 2018 20:53:35 +0200 Subject: net/mlx5: Remove not-used lib/eq.h header file lib/eq.h is needed for EQ manipulation which are not performed in SRQ. Reviewed-by: Mark Bloch Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/srq.c | 1 - 1 file changed, 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/srq.c b/drivers/net/ethernet/mellanox/mlx5/core/srq.c index 10036aaa200a..248f1c8dd5d3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/srq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/srq.c @@ -11,7 +11,6 @@ #include #include #include "mlx5_core.h" -#include "lib/eq.h" static int srq_event_notifier(struct notifier_block *nb, unsigned long type, void *data) -- cgit v1.2.3-55-g7522 From c23f88cb575fe4a0b49603d7f9e5c3782886269f Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Wed, 28 Nov 2018 20:53:36 +0200 Subject: net/mlx5: Remove references to local mlx5_core functions As a preparation to move SRQ functionality to RDMA, drop all references to mlx5_core logic and make SRQ be dependent on shared code only. Most of the time, we are interested to know if events are working/not working and it is possible with previous commit ("net/mlx5: Debug print for forwarded async events"). Reviewed-by: Mark Bloch Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/srq.c | 22 +++------------------- 1 file changed, 3 insertions(+), 19 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/srq.c b/drivers/net/ethernet/mellanox/mlx5/core/srq.c index 248f1c8dd5d3..690815234838 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/srq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/srq.c @@ -10,15 +10,12 @@ #include #include #include -#include "mlx5_core.h" static int srq_event_notifier(struct notifier_block *nb, unsigned long type, void *data) { struct mlx5_srq_table *table; - struct mlx5_core_dev *dev; struct mlx5_core_srq *srq; - struct mlx5_priv *priv; struct mlx5_eqe *eqe; u32 srqn; @@ -27,12 +24,9 @@ static int srq_event_notifier(struct notifier_block *nb, return NOTIFY_DONE; table = container_of(nb, struct mlx5_srq_table, nb); - priv = container_of(table, struct mlx5_priv, srq_table); - dev = container_of(priv, struct mlx5_core_dev, priv); eqe = data; srqn = be32_to_cpu(eqe->data.qp_srq.qp_srq_n) & 0xffffff; - mlx5_core_dbg(dev, "SRQ event (%d): srqn 0x%x\n", eqe->type, srqn); spin_lock(&table->lock); @@ -42,10 +36,8 @@ static int srq_event_notifier(struct notifier_block *nb, spin_unlock(&table->lock); - if (!srq) { - mlx5_core_warn(dev, "Async event for bogus SRQ 0x%08x\n", srqn); + if (!srq) return NOTIFY_OK; - } srq->event(srq, eqe->type); @@ -617,10 +609,8 @@ int mlx5_core_create_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, spin_lock_irq(&table->lock); err = radix_tree_insert(&table->tree, srq->srqn, srq); spin_unlock_irq(&table->lock); - if (err) { - mlx5_core_warn(dev, "err %d, srqn 0x%x\n", err, srq->srqn); + if (err) goto err_destroy_srq_split; - } return 0; @@ -640,14 +630,8 @@ int mlx5_core_destroy_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq) spin_lock_irq(&table->lock); tmp = radix_tree_delete(&table->tree, srq->srqn); spin_unlock_irq(&table->lock); - if (!tmp) { - mlx5_core_warn(dev, "srq 0x%x not found in tree\n", srq->srqn); + if (!tmp || tmp != srq) return -EINVAL; - } - if (tmp != srq) { - mlx5_core_warn(dev, "corruption on srqn 0x%x\n", srq->srqn); - return -EINVAL; - } err = destroy_srq_split(dev, srq); if (err) -- cgit v1.2.3-55-g7522 From f02d0d6e53ac2c8a75b6cc87dc86675a9351d84d Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Wed, 28 Nov 2018 20:53:37 +0200 Subject: net/mlx5: Move SRQ functions to RDMA part There is no need to keep SRQ which is RDMA object in mlx5_core. In this patch, we partially move the execution code, while next patches will move table initialization/release logic too. Reviewed-by: Mark Bloch Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/mlx5/Makefile | 4 +- drivers/infiniband/hw/mlx5/cq.c | 1 + drivers/infiniband/hw/mlx5/srq.c | 2 +- drivers/infiniband/hw/mlx5/srq.h | 46 ++ drivers/infiniband/hw/mlx5/srq_cmd.c | 666 +++++++++++++++++++++ drivers/net/ethernet/mellanox/mlx5/core/srq.c | 634 -------------------- drivers/net/ethernet/mellanox/mlx5/core/transobj.c | 43 -- include/linux/mlx5/driver.h | 8 - include/linux/mlx5/srq.h | 31 - include/linux/mlx5/transobj.h | 6 - 10 files changed, 717 insertions(+), 724 deletions(-) create mode 100644 drivers/infiniband/hw/mlx5/srq.h create mode 100644 drivers/infiniband/hw/mlx5/srq_cmd.c (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/infiniband/hw/mlx5/Makefile b/drivers/infiniband/hw/mlx5/Makefile index b8e4b15e2674..33f5adb14e4e 100644 --- a/drivers/infiniband/hw/mlx5/Makefile +++ b/drivers/infiniband/hw/mlx5/Makefile @@ -1,6 +1,8 @@ obj-$(CONFIG_MLX5_INFINIBAND) += mlx5_ib.o -mlx5_ib-y := main.o cq.o doorbell.o qp.o mem.o srq.o mr.o ah.o mad.o gsi.o ib_virt.o cmd.o cong.o +mlx5_ib-y := main.o cq.o doorbell.o qp.o mem.o srq_cmd.o \ + srq.o mr.o ah.o mad.o gsi.o ib_virt.o cmd.o \ + cong.o mlx5_ib-$(CONFIG_INFINIBAND_ON_DEMAND_PAGING) += odp.o mlx5_ib-$(CONFIG_MLX5_ESWITCH) += ib_rep.o mlx5_ib-$(CONFIG_INFINIBAND_USER_ACCESS) += devx.o diff --git a/drivers/infiniband/hw/mlx5/cq.c b/drivers/infiniband/hw/mlx5/cq.c index 7d769b5538b4..c5d2824ada59 100644 --- a/drivers/infiniband/hw/mlx5/cq.c +++ b/drivers/infiniband/hw/mlx5/cq.c @@ -35,6 +35,7 @@ #include #include #include "mlx5_ib.h" +#include "srq.h" static void mlx5_ib_cq_comp(struct mlx5_core_cq *cq) { diff --git a/drivers/infiniband/hw/mlx5/srq.c b/drivers/infiniband/hw/mlx5/srq.c index 28794780062e..a86d9f153805 100644 --- a/drivers/infiniband/hw/mlx5/srq.c +++ b/drivers/infiniband/hw/mlx5/srq.c @@ -9,8 +9,8 @@ #include #include #include - #include "mlx5_ib.h" +#include "srq.h" /* not supported currently */ static int srq_signature; diff --git a/drivers/infiniband/hw/mlx5/srq.h b/drivers/infiniband/hw/mlx5/srq.h new file mode 100644 index 000000000000..f23d5de12973 --- /dev/null +++ b/drivers/infiniband/hw/mlx5/srq.h @@ -0,0 +1,46 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* + * Copyright (c) 2013-2018, Mellanox Technologies. All rights reserved. + */ + +#ifndef MLX5_IB_SRQ_H +#define MLX5_IB_SRQ_H + +enum { + MLX5_SRQ_FLAG_ERR = (1 << 0), + MLX5_SRQ_FLAG_WQ_SIG = (1 << 1), + MLX5_SRQ_FLAG_RNDV = (1 << 2), +}; + +struct mlx5_srq_attr { + u32 type; + u32 flags; + u32 log_size; + u32 wqe_shift; + u32 log_page_size; + u32 wqe_cnt; + u32 srqn; + u32 xrcd; + u32 page_offset; + u32 cqn; + u32 pd; + u32 lwm; + u32 user_index; + u64 db_record; + __be64 *pas; + u32 tm_log_list_size; + u32 tm_next_tag; + u32 tm_hw_phase_cnt; + u32 tm_sw_phase_cnt; + u16 uid; +}; + +int mlx5_core_create_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + struct mlx5_srq_attr *in); +int mlx5_core_destroy_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq); +int mlx5_core_query_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + struct mlx5_srq_attr *out); +int mlx5_core_arm_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + u16 lwm, int is_srq); +struct mlx5_core_srq *mlx5_core_get_srq(struct mlx5_core_dev *dev, u32 srqn); +#endif /* MLX5_IB_SRQ_H */ diff --git a/drivers/infiniband/hw/mlx5/srq_cmd.c b/drivers/infiniband/hw/mlx5/srq_cmd.c new file mode 100644 index 000000000000..4a64ad4c9b25 --- /dev/null +++ b/drivers/infiniband/hw/mlx5/srq_cmd.c @@ -0,0 +1,666 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB +/* + * Copyright (c) 2013-2018, Mellanox Technologies inc. All rights reserved. + */ + +#include +#include +#include +#include "srq.h" + +static int get_pas_size(struct mlx5_srq_attr *in) +{ + u32 log_page_size = in->log_page_size + 12; + u32 log_srq_size = in->log_size; + u32 log_rq_stride = in->wqe_shift; + u32 page_offset = in->page_offset; + u32 po_quanta = 1 << (log_page_size - 6); + u32 rq_sz = 1 << (log_srq_size + 4 + log_rq_stride); + u32 page_size = 1 << log_page_size; + u32 rq_sz_po = rq_sz + (page_offset * po_quanta); + u32 rq_num_pas = DIV_ROUND_UP(rq_sz_po, page_size); + + return rq_num_pas * sizeof(u64); +} + +static void set_wq(void *wq, struct mlx5_srq_attr *in) +{ + MLX5_SET(wq, wq, wq_signature, !!(in->flags + & MLX5_SRQ_FLAG_WQ_SIG)); + MLX5_SET(wq, wq, log_wq_pg_sz, in->log_page_size); + MLX5_SET(wq, wq, log_wq_stride, in->wqe_shift + 4); + MLX5_SET(wq, wq, log_wq_sz, in->log_size); + MLX5_SET(wq, wq, page_offset, in->page_offset); + MLX5_SET(wq, wq, lwm, in->lwm); + MLX5_SET(wq, wq, pd, in->pd); + MLX5_SET64(wq, wq, dbr_addr, in->db_record); +} + +static void set_srqc(void *srqc, struct mlx5_srq_attr *in) +{ + MLX5_SET(srqc, srqc, wq_signature, !!(in->flags + & MLX5_SRQ_FLAG_WQ_SIG)); + MLX5_SET(srqc, srqc, log_page_size, in->log_page_size); + MLX5_SET(srqc, srqc, log_rq_stride, in->wqe_shift); + MLX5_SET(srqc, srqc, log_srq_size, in->log_size); + MLX5_SET(srqc, srqc, page_offset, in->page_offset); + MLX5_SET(srqc, srqc, lwm, in->lwm); + MLX5_SET(srqc, srqc, pd, in->pd); + MLX5_SET64(srqc, srqc, dbr_addr, in->db_record); + MLX5_SET(srqc, srqc, xrcd, in->xrcd); + MLX5_SET(srqc, srqc, cqn, in->cqn); +} + +static void get_wq(void *wq, struct mlx5_srq_attr *in) +{ + if (MLX5_GET(wq, wq, wq_signature)) + in->flags &= MLX5_SRQ_FLAG_WQ_SIG; + in->log_page_size = MLX5_GET(wq, wq, log_wq_pg_sz); + in->wqe_shift = MLX5_GET(wq, wq, log_wq_stride) - 4; + in->log_size = MLX5_GET(wq, wq, log_wq_sz); + in->page_offset = MLX5_GET(wq, wq, page_offset); + in->lwm = MLX5_GET(wq, wq, lwm); + in->pd = MLX5_GET(wq, wq, pd); + in->db_record = MLX5_GET64(wq, wq, dbr_addr); +} + +static void get_srqc(void *srqc, struct mlx5_srq_attr *in) +{ + if (MLX5_GET(srqc, srqc, wq_signature)) + in->flags &= MLX5_SRQ_FLAG_WQ_SIG; + in->log_page_size = MLX5_GET(srqc, srqc, log_page_size); + in->wqe_shift = MLX5_GET(srqc, srqc, log_rq_stride); + in->log_size = MLX5_GET(srqc, srqc, log_srq_size); + in->page_offset = MLX5_GET(srqc, srqc, page_offset); + in->lwm = MLX5_GET(srqc, srqc, lwm); + in->pd = MLX5_GET(srqc, srqc, pd); + in->db_record = MLX5_GET64(srqc, srqc, dbr_addr); +} + +struct mlx5_core_srq *mlx5_core_get_srq(struct mlx5_core_dev *dev, u32 srqn) +{ + struct mlx5_srq_table *table = &dev->priv.srq_table; + struct mlx5_core_srq *srq; + + spin_lock(&table->lock); + + srq = radix_tree_lookup(&table->tree, srqn); + if (srq) + atomic_inc(&srq->refcount); + + spin_unlock(&table->lock); + + return srq; +} + +static int create_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + struct mlx5_srq_attr *in) +{ + u32 create_out[MLX5_ST_SZ_DW(create_srq_out)] = {0}; + void *create_in; + void *srqc; + void *pas; + int pas_size; + int inlen; + int err; + + pas_size = get_pas_size(in); + inlen = MLX5_ST_SZ_BYTES(create_srq_in) + pas_size; + create_in = kvzalloc(inlen, GFP_KERNEL); + if (!create_in) + return -ENOMEM; + + MLX5_SET(create_srq_in, create_in, uid, in->uid); + srqc = MLX5_ADDR_OF(create_srq_in, create_in, srq_context_entry); + pas = MLX5_ADDR_OF(create_srq_in, create_in, pas); + + set_srqc(srqc, in); + memcpy(pas, in->pas, pas_size); + + MLX5_SET(create_srq_in, create_in, opcode, + MLX5_CMD_OP_CREATE_SRQ); + + err = mlx5_cmd_exec(dev, create_in, inlen, create_out, + sizeof(create_out)); + kvfree(create_in); + if (!err) { + srq->srqn = MLX5_GET(create_srq_out, create_out, srqn); + srq->uid = in->uid; + } + + return err; +} + +static int destroy_srq_cmd(struct mlx5_core_dev *dev, + struct mlx5_core_srq *srq) +{ + u32 srq_in[MLX5_ST_SZ_DW(destroy_srq_in)] = {0}; + u32 srq_out[MLX5_ST_SZ_DW(destroy_srq_out)] = {0}; + + MLX5_SET(destroy_srq_in, srq_in, opcode, + MLX5_CMD_OP_DESTROY_SRQ); + MLX5_SET(destroy_srq_in, srq_in, srqn, srq->srqn); + MLX5_SET(destroy_srq_in, srq_in, uid, srq->uid); + + return mlx5_cmd_exec(dev, srq_in, sizeof(srq_in), + srq_out, sizeof(srq_out)); +} + +static int arm_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + u16 lwm, int is_srq) +{ + u32 srq_in[MLX5_ST_SZ_DW(arm_rq_in)] = {0}; + u32 srq_out[MLX5_ST_SZ_DW(arm_rq_out)] = {0}; + + MLX5_SET(arm_rq_in, srq_in, opcode, MLX5_CMD_OP_ARM_RQ); + MLX5_SET(arm_rq_in, srq_in, op_mod, MLX5_ARM_RQ_IN_OP_MOD_SRQ); + MLX5_SET(arm_rq_in, srq_in, srq_number, srq->srqn); + MLX5_SET(arm_rq_in, srq_in, lwm, lwm); + MLX5_SET(arm_rq_in, srq_in, uid, srq->uid); + + return mlx5_cmd_exec(dev, srq_in, sizeof(srq_in), + srq_out, sizeof(srq_out)); +} + +static int query_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + struct mlx5_srq_attr *out) +{ + u32 srq_in[MLX5_ST_SZ_DW(query_srq_in)] = {0}; + u32 *srq_out; + void *srqc; + int err; + + srq_out = kvzalloc(MLX5_ST_SZ_BYTES(query_srq_out), GFP_KERNEL); + if (!srq_out) + return -ENOMEM; + + MLX5_SET(query_srq_in, srq_in, opcode, + MLX5_CMD_OP_QUERY_SRQ); + MLX5_SET(query_srq_in, srq_in, srqn, srq->srqn); + err = mlx5_cmd_exec(dev, srq_in, sizeof(srq_in), + srq_out, MLX5_ST_SZ_BYTES(query_srq_out)); + if (err) + goto out; + + srqc = MLX5_ADDR_OF(query_srq_out, srq_out, srq_context_entry); + get_srqc(srqc, out); + if (MLX5_GET(srqc, srqc, state) != MLX5_SRQC_STATE_GOOD) + out->flags |= MLX5_SRQ_FLAG_ERR; +out: + kvfree(srq_out); + return err; +} + +static int create_xrc_srq_cmd(struct mlx5_core_dev *dev, + struct mlx5_core_srq *srq, + struct mlx5_srq_attr *in) +{ + u32 create_out[MLX5_ST_SZ_DW(create_xrc_srq_out)]; + void *create_in; + void *xrc_srqc; + void *pas; + int pas_size; + int inlen; + int err; + + pas_size = get_pas_size(in); + inlen = MLX5_ST_SZ_BYTES(create_xrc_srq_in) + pas_size; + create_in = kvzalloc(inlen, GFP_KERNEL); + if (!create_in) + return -ENOMEM; + + MLX5_SET(create_xrc_srq_in, create_in, uid, in->uid); + xrc_srqc = MLX5_ADDR_OF(create_xrc_srq_in, create_in, + xrc_srq_context_entry); + pas = MLX5_ADDR_OF(create_xrc_srq_in, create_in, pas); + + set_srqc(xrc_srqc, in); + MLX5_SET(xrc_srqc, xrc_srqc, user_index, in->user_index); + memcpy(pas, in->pas, pas_size); + MLX5_SET(create_xrc_srq_in, create_in, opcode, + MLX5_CMD_OP_CREATE_XRC_SRQ); + + memset(create_out, 0, sizeof(create_out)); + err = mlx5_cmd_exec(dev, create_in, inlen, create_out, + sizeof(create_out)); + if (err) + goto out; + + srq->srqn = MLX5_GET(create_xrc_srq_out, create_out, xrc_srqn); + srq->uid = in->uid; +out: + kvfree(create_in); + return err; +} + +static int destroy_xrc_srq_cmd(struct mlx5_core_dev *dev, + struct mlx5_core_srq *srq) +{ + u32 xrcsrq_in[MLX5_ST_SZ_DW(destroy_xrc_srq_in)] = {0}; + u32 xrcsrq_out[MLX5_ST_SZ_DW(destroy_xrc_srq_out)] = {0}; + + MLX5_SET(destroy_xrc_srq_in, xrcsrq_in, opcode, + MLX5_CMD_OP_DESTROY_XRC_SRQ); + MLX5_SET(destroy_xrc_srq_in, xrcsrq_in, xrc_srqn, srq->srqn); + MLX5_SET(destroy_xrc_srq_in, xrcsrq_in, uid, srq->uid); + + return mlx5_cmd_exec(dev, xrcsrq_in, sizeof(xrcsrq_in), + xrcsrq_out, sizeof(xrcsrq_out)); +} + +static int arm_xrc_srq_cmd(struct mlx5_core_dev *dev, + struct mlx5_core_srq *srq, u16 lwm) +{ + u32 xrcsrq_in[MLX5_ST_SZ_DW(arm_xrc_srq_in)] = {0}; + u32 xrcsrq_out[MLX5_ST_SZ_DW(arm_xrc_srq_out)] = {0}; + + MLX5_SET(arm_xrc_srq_in, xrcsrq_in, opcode, MLX5_CMD_OP_ARM_XRC_SRQ); + MLX5_SET(arm_xrc_srq_in, xrcsrq_in, op_mod, MLX5_ARM_XRC_SRQ_IN_OP_MOD_XRC_SRQ); + MLX5_SET(arm_xrc_srq_in, xrcsrq_in, xrc_srqn, srq->srqn); + MLX5_SET(arm_xrc_srq_in, xrcsrq_in, lwm, lwm); + MLX5_SET(arm_xrc_srq_in, xrcsrq_in, uid, srq->uid); + + return mlx5_cmd_exec(dev, xrcsrq_in, sizeof(xrcsrq_in), + xrcsrq_out, sizeof(xrcsrq_out)); +} + +static int query_xrc_srq_cmd(struct mlx5_core_dev *dev, + struct mlx5_core_srq *srq, + struct mlx5_srq_attr *out) +{ + u32 xrcsrq_in[MLX5_ST_SZ_DW(query_xrc_srq_in)]; + u32 *xrcsrq_out; + void *xrc_srqc; + int err; + + xrcsrq_out = kvzalloc(MLX5_ST_SZ_BYTES(query_xrc_srq_out), GFP_KERNEL); + if (!xrcsrq_out) + return -ENOMEM; + memset(xrcsrq_in, 0, sizeof(xrcsrq_in)); + + MLX5_SET(query_xrc_srq_in, xrcsrq_in, opcode, + MLX5_CMD_OP_QUERY_XRC_SRQ); + MLX5_SET(query_xrc_srq_in, xrcsrq_in, xrc_srqn, srq->srqn); + + err = mlx5_cmd_exec(dev, xrcsrq_in, sizeof(xrcsrq_in), xrcsrq_out, + MLX5_ST_SZ_BYTES(query_xrc_srq_out)); + if (err) + goto out; + + xrc_srqc = MLX5_ADDR_OF(query_xrc_srq_out, xrcsrq_out, + xrc_srq_context_entry); + get_srqc(xrc_srqc, out); + if (MLX5_GET(xrc_srqc, xrc_srqc, state) != MLX5_XRC_SRQC_STATE_GOOD) + out->flags |= MLX5_SRQ_FLAG_ERR; + +out: + kvfree(xrcsrq_out); + return err; +} + +static int mlx5_core_create_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen, + u32 *rmpn) +{ + u32 out[MLX5_ST_SZ_DW(create_rmp_out)] = { 0 }; + int err; + + MLX5_SET(create_rmp_in, in, opcode, MLX5_CMD_OP_CREATE_RMP); + err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out)); + if (!err) + *rmpn = MLX5_GET(create_rmp_out, out, rmpn); + + return err; +} + +static int mlx5_core_modify_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen) +{ + u32 out[MLX5_ST_SZ_DW(modify_rmp_out)] = {0}; + + MLX5_SET(modify_rmp_in, in, opcode, MLX5_CMD_OP_MODIFY_RMP); + return mlx5_cmd_exec(dev, in, inlen, out, sizeof(out)); +} + +static int mlx5_core_query_rmp(struct mlx5_core_dev *dev, u32 rmpn, u32 *out) +{ + u32 in[MLX5_ST_SZ_DW(query_rmp_in)] = {0}; + int outlen = MLX5_ST_SZ_BYTES(query_rmp_out); + + MLX5_SET(query_rmp_in, in, opcode, MLX5_CMD_OP_QUERY_RMP); + MLX5_SET(query_rmp_in, in, rmpn, rmpn); + return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen); +} + +static int create_rmp_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + struct mlx5_srq_attr *in) +{ + void *create_in; + void *rmpc; + void *wq; + int pas_size; + int inlen; + int err; + + pas_size = get_pas_size(in); + inlen = MLX5_ST_SZ_BYTES(create_rmp_in) + pas_size; + create_in = kvzalloc(inlen, GFP_KERNEL); + if (!create_in) + return -ENOMEM; + + rmpc = MLX5_ADDR_OF(create_rmp_in, create_in, ctx); + wq = MLX5_ADDR_OF(rmpc, rmpc, wq); + + MLX5_SET(rmpc, rmpc, state, MLX5_RMPC_STATE_RDY); + MLX5_SET(create_rmp_in, create_in, uid, in->uid); + set_wq(wq, in); + memcpy(MLX5_ADDR_OF(rmpc, rmpc, wq.pas), in->pas, pas_size); + + err = mlx5_core_create_rmp(dev, create_in, inlen, &srq->srqn); + if (!err) + srq->uid = in->uid; + + kvfree(create_in); + return err; +} + +static int destroy_rmp_cmd(struct mlx5_core_dev *dev, + struct mlx5_core_srq *srq) +{ + u32 in[MLX5_ST_SZ_DW(destroy_rmp_in)] = {}; + u32 out[MLX5_ST_SZ_DW(destroy_rmp_out)] = {}; + + MLX5_SET(destroy_rmp_in, in, opcode, MLX5_CMD_OP_DESTROY_RMP); + MLX5_SET(destroy_rmp_in, in, rmpn, srq->srqn); + MLX5_SET(destroy_rmp_in, in, uid, srq->uid); + return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); +} + +static int arm_rmp_cmd(struct mlx5_core_dev *dev, + struct mlx5_core_srq *srq, + u16 lwm) +{ + void *in; + void *rmpc; + void *wq; + void *bitmask; + int err; + + in = kvzalloc(MLX5_ST_SZ_BYTES(modify_rmp_in), GFP_KERNEL); + if (!in) + return -ENOMEM; + + rmpc = MLX5_ADDR_OF(modify_rmp_in, in, ctx); + bitmask = MLX5_ADDR_OF(modify_rmp_in, in, bitmask); + wq = MLX5_ADDR_OF(rmpc, rmpc, wq); + + MLX5_SET(modify_rmp_in, in, rmp_state, MLX5_RMPC_STATE_RDY); + MLX5_SET(modify_rmp_in, in, rmpn, srq->srqn); + MLX5_SET(modify_rmp_in, in, uid, srq->uid); + MLX5_SET(wq, wq, lwm, lwm); + MLX5_SET(rmp_bitmask, bitmask, lwm, 1); + MLX5_SET(rmpc, rmpc, state, MLX5_RMPC_STATE_RDY); + + err = mlx5_core_modify_rmp(dev, in, MLX5_ST_SZ_BYTES(modify_rmp_in)); + + kvfree(in); + return err; +} + +static int query_rmp_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + struct mlx5_srq_attr *out) +{ + u32 *rmp_out; + void *rmpc; + int err; + + rmp_out = kvzalloc(MLX5_ST_SZ_BYTES(query_rmp_out), GFP_KERNEL); + if (!rmp_out) + return -ENOMEM; + + err = mlx5_core_query_rmp(dev, srq->srqn, rmp_out); + if (err) + goto out; + + rmpc = MLX5_ADDR_OF(query_rmp_out, rmp_out, rmp_context); + get_wq(MLX5_ADDR_OF(rmpc, rmpc, wq), out); + if (MLX5_GET(rmpc, rmpc, state) != MLX5_RMPC_STATE_RDY) + out->flags |= MLX5_SRQ_FLAG_ERR; + +out: + kvfree(rmp_out); + return err; +} + +static int create_xrq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + struct mlx5_srq_attr *in) +{ + u32 create_out[MLX5_ST_SZ_DW(create_xrq_out)] = {0}; + void *create_in; + void *xrqc; + void *wq; + int pas_size; + int inlen; + int err; + + pas_size = get_pas_size(in); + inlen = MLX5_ST_SZ_BYTES(create_xrq_in) + pas_size; + create_in = kvzalloc(inlen, GFP_KERNEL); + if (!create_in) + return -ENOMEM; + + xrqc = MLX5_ADDR_OF(create_xrq_in, create_in, xrq_context); + wq = MLX5_ADDR_OF(xrqc, xrqc, wq); + + set_wq(wq, in); + memcpy(MLX5_ADDR_OF(xrqc, xrqc, wq.pas), in->pas, pas_size); + + if (in->type == IB_SRQT_TM) { + MLX5_SET(xrqc, xrqc, topology, MLX5_XRQC_TOPOLOGY_TAG_MATCHING); + if (in->flags & MLX5_SRQ_FLAG_RNDV) + MLX5_SET(xrqc, xrqc, offload, MLX5_XRQC_OFFLOAD_RNDV); + MLX5_SET(xrqc, xrqc, + tag_matching_topology_context.log_matching_list_sz, + in->tm_log_list_size); + } + MLX5_SET(xrqc, xrqc, user_index, in->user_index); + MLX5_SET(xrqc, xrqc, cqn, in->cqn); + MLX5_SET(create_xrq_in, create_in, opcode, MLX5_CMD_OP_CREATE_XRQ); + MLX5_SET(create_xrq_in, create_in, uid, in->uid); + err = mlx5_cmd_exec(dev, create_in, inlen, create_out, + sizeof(create_out)); + kvfree(create_in); + if (!err) { + srq->srqn = MLX5_GET(create_xrq_out, create_out, xrqn); + srq->uid = in->uid; + } + + return err; +} + +static int destroy_xrq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq) +{ + u32 in[MLX5_ST_SZ_DW(destroy_xrq_in)] = {0}; + u32 out[MLX5_ST_SZ_DW(destroy_xrq_out)] = {0}; + + MLX5_SET(destroy_xrq_in, in, opcode, MLX5_CMD_OP_DESTROY_XRQ); + MLX5_SET(destroy_xrq_in, in, xrqn, srq->srqn); + MLX5_SET(destroy_xrq_in, in, uid, srq->uid); + + return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); +} + +static int arm_xrq_cmd(struct mlx5_core_dev *dev, + struct mlx5_core_srq *srq, + u16 lwm) +{ + u32 out[MLX5_ST_SZ_DW(arm_rq_out)] = {0}; + u32 in[MLX5_ST_SZ_DW(arm_rq_in)] = {0}; + + MLX5_SET(arm_rq_in, in, opcode, MLX5_CMD_OP_ARM_RQ); + MLX5_SET(arm_rq_in, in, op_mod, MLX5_ARM_RQ_IN_OP_MOD_XRQ); + MLX5_SET(arm_rq_in, in, srq_number, srq->srqn); + MLX5_SET(arm_rq_in, in, lwm, lwm); + MLX5_SET(arm_rq_in, in, uid, srq->uid); + + return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); +} + +static int query_xrq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + struct mlx5_srq_attr *out) +{ + u32 in[MLX5_ST_SZ_DW(query_xrq_in)] = {0}; + u32 *xrq_out; + int outlen = MLX5_ST_SZ_BYTES(query_xrq_out); + void *xrqc; + int err; + + xrq_out = kvzalloc(outlen, GFP_KERNEL); + if (!xrq_out) + return -ENOMEM; + + MLX5_SET(query_xrq_in, in, opcode, MLX5_CMD_OP_QUERY_XRQ); + MLX5_SET(query_xrq_in, in, xrqn, srq->srqn); + + err = mlx5_cmd_exec(dev, in, sizeof(in), xrq_out, outlen); + if (err) + goto out; + + xrqc = MLX5_ADDR_OF(query_xrq_out, xrq_out, xrq_context); + get_wq(MLX5_ADDR_OF(xrqc, xrqc, wq), out); + if (MLX5_GET(xrqc, xrqc, state) != MLX5_XRQC_STATE_GOOD) + out->flags |= MLX5_SRQ_FLAG_ERR; + out->tm_next_tag = + MLX5_GET(xrqc, xrqc, + tag_matching_topology_context.append_next_index); + out->tm_hw_phase_cnt = + MLX5_GET(xrqc, xrqc, + tag_matching_topology_context.hw_phase_cnt); + out->tm_sw_phase_cnt = + MLX5_GET(xrqc, xrqc, + tag_matching_topology_context.sw_phase_cnt); + +out: + kvfree(xrq_out); + return err; +} + +static int create_srq_split(struct mlx5_core_dev *dev, + struct mlx5_core_srq *srq, + struct mlx5_srq_attr *in) +{ + if (!dev->issi) + return create_srq_cmd(dev, srq, in); + switch (srq->common.res) { + case MLX5_RES_XSRQ: + return create_xrc_srq_cmd(dev, srq, in); + case MLX5_RES_XRQ: + return create_xrq_cmd(dev, srq, in); + default: + return create_rmp_cmd(dev, srq, in); + } +} + +static int destroy_srq_split(struct mlx5_core_dev *dev, + struct mlx5_core_srq *srq) +{ + if (!dev->issi) + return destroy_srq_cmd(dev, srq); + switch (srq->common.res) { + case MLX5_RES_XSRQ: + return destroy_xrc_srq_cmd(dev, srq); + case MLX5_RES_XRQ: + return destroy_xrq_cmd(dev, srq); + default: + return destroy_rmp_cmd(dev, srq); + } +} + +int mlx5_core_create_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + struct mlx5_srq_attr *in) +{ + int err; + struct mlx5_srq_table *table = &dev->priv.srq_table; + + switch (in->type) { + case IB_SRQT_XRC: + srq->common.res = MLX5_RES_XSRQ; + break; + case IB_SRQT_TM: + srq->common.res = MLX5_RES_XRQ; + break; + default: + srq->common.res = MLX5_RES_SRQ; + } + + err = create_srq_split(dev, srq, in); + if (err) + return err; + + atomic_set(&srq->refcount, 1); + init_completion(&srq->free); + + spin_lock_irq(&table->lock); + err = radix_tree_insert(&table->tree, srq->srqn, srq); + spin_unlock_irq(&table->lock); + if (err) + goto err_destroy_srq_split; + + return 0; + +err_destroy_srq_split: + destroy_srq_split(dev, srq); + + return err; +} + +int mlx5_core_destroy_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq) +{ + struct mlx5_srq_table *table = &dev->priv.srq_table; + struct mlx5_core_srq *tmp; + int err; + + spin_lock_irq(&table->lock); + tmp = radix_tree_delete(&table->tree, srq->srqn); + spin_unlock_irq(&table->lock); + if (!tmp || tmp != srq) + return -EINVAL; + + err = destroy_srq_split(dev, srq); + if (err) + return err; + + if (atomic_dec_and_test(&srq->refcount)) + complete(&srq->free); + wait_for_completion(&srq->free); + + return 0; +} + +int mlx5_core_query_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + struct mlx5_srq_attr *out) +{ + if (!dev->issi) + return query_srq_cmd(dev, srq, out); + switch (srq->common.res) { + case MLX5_RES_XSRQ: + return query_xrc_srq_cmd(dev, srq, out); + case MLX5_RES_XRQ: + return query_xrq_cmd(dev, srq, out); + default: + return query_rmp_cmd(dev, srq, out); + } +} + +int mlx5_core_arm_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, + u16 lwm, int is_srq) +{ + if (!dev->issi) + return arm_srq_cmd(dev, srq, lwm, is_srq); + switch (srq->common.res) { + case MLX5_RES_XSRQ: + return arm_xrc_srq_cmd(dev, srq, lwm); + case MLX5_RES_XRQ: + return arm_xrq_cmd(dev, srq, lwm); + default: + return arm_rmp_cmd(dev, srq, lwm); + } +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/srq.c b/drivers/net/ethernet/mellanox/mlx5/core/srq.c index 690815234838..0e80ddbe2510 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/srq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/srq.c @@ -4,12 +4,8 @@ */ #include -#include #include -#include #include -#include -#include static int srq_event_notifier(struct notifier_block *nb, unsigned long type, void *data) @@ -47,636 +43,6 @@ static int srq_event_notifier(struct notifier_block *nb, return NOTIFY_OK; } -static int get_pas_size(struct mlx5_srq_attr *in) -{ - u32 log_page_size = in->log_page_size + 12; - u32 log_srq_size = in->log_size; - u32 log_rq_stride = in->wqe_shift; - u32 page_offset = in->page_offset; - u32 po_quanta = 1 << (log_page_size - 6); - u32 rq_sz = 1 << (log_srq_size + 4 + log_rq_stride); - u32 page_size = 1 << log_page_size; - u32 rq_sz_po = rq_sz + (page_offset * po_quanta); - u32 rq_num_pas = DIV_ROUND_UP(rq_sz_po, page_size); - - return rq_num_pas * sizeof(u64); -} - -static void set_wq(void *wq, struct mlx5_srq_attr *in) -{ - MLX5_SET(wq, wq, wq_signature, !!(in->flags - & MLX5_SRQ_FLAG_WQ_SIG)); - MLX5_SET(wq, wq, log_wq_pg_sz, in->log_page_size); - MLX5_SET(wq, wq, log_wq_stride, in->wqe_shift + 4); - MLX5_SET(wq, wq, log_wq_sz, in->log_size); - MLX5_SET(wq, wq, page_offset, in->page_offset); - MLX5_SET(wq, wq, lwm, in->lwm); - MLX5_SET(wq, wq, pd, in->pd); - MLX5_SET64(wq, wq, dbr_addr, in->db_record); -} - -static void set_srqc(void *srqc, struct mlx5_srq_attr *in) -{ - MLX5_SET(srqc, srqc, wq_signature, !!(in->flags - & MLX5_SRQ_FLAG_WQ_SIG)); - MLX5_SET(srqc, srqc, log_page_size, in->log_page_size); - MLX5_SET(srqc, srqc, log_rq_stride, in->wqe_shift); - MLX5_SET(srqc, srqc, log_srq_size, in->log_size); - MLX5_SET(srqc, srqc, page_offset, in->page_offset); - MLX5_SET(srqc, srqc, lwm, in->lwm); - MLX5_SET(srqc, srqc, pd, in->pd); - MLX5_SET64(srqc, srqc, dbr_addr, in->db_record); - MLX5_SET(srqc, srqc, xrcd, in->xrcd); - MLX5_SET(srqc, srqc, cqn, in->cqn); -} - -static void get_wq(void *wq, struct mlx5_srq_attr *in) -{ - if (MLX5_GET(wq, wq, wq_signature)) - in->flags &= MLX5_SRQ_FLAG_WQ_SIG; - in->log_page_size = MLX5_GET(wq, wq, log_wq_pg_sz); - in->wqe_shift = MLX5_GET(wq, wq, log_wq_stride) - 4; - in->log_size = MLX5_GET(wq, wq, log_wq_sz); - in->page_offset = MLX5_GET(wq, wq, page_offset); - in->lwm = MLX5_GET(wq, wq, lwm); - in->pd = MLX5_GET(wq, wq, pd); - in->db_record = MLX5_GET64(wq, wq, dbr_addr); -} - -static void get_srqc(void *srqc, struct mlx5_srq_attr *in) -{ - if (MLX5_GET(srqc, srqc, wq_signature)) - in->flags &= MLX5_SRQ_FLAG_WQ_SIG; - in->log_page_size = MLX5_GET(srqc, srqc, log_page_size); - in->wqe_shift = MLX5_GET(srqc, srqc, log_rq_stride); - in->log_size = MLX5_GET(srqc, srqc, log_srq_size); - in->page_offset = MLX5_GET(srqc, srqc, page_offset); - in->lwm = MLX5_GET(srqc, srqc, lwm); - in->pd = MLX5_GET(srqc, srqc, pd); - in->db_record = MLX5_GET64(srqc, srqc, dbr_addr); -} - -struct mlx5_core_srq *mlx5_core_get_srq(struct mlx5_core_dev *dev, u32 srqn) -{ - struct mlx5_srq_table *table = &dev->priv.srq_table; - struct mlx5_core_srq *srq; - - spin_lock(&table->lock); - - srq = radix_tree_lookup(&table->tree, srqn); - if (srq) - atomic_inc(&srq->refcount); - - spin_unlock(&table->lock); - - return srq; -} -EXPORT_SYMBOL(mlx5_core_get_srq); - -static int create_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - struct mlx5_srq_attr *in) -{ - u32 create_out[MLX5_ST_SZ_DW(create_srq_out)] = {0}; - void *create_in; - void *srqc; - void *pas; - int pas_size; - int inlen; - int err; - - pas_size = get_pas_size(in); - inlen = MLX5_ST_SZ_BYTES(create_srq_in) + pas_size; - create_in = kvzalloc(inlen, GFP_KERNEL); - if (!create_in) - return -ENOMEM; - - MLX5_SET(create_srq_in, create_in, uid, in->uid); - srqc = MLX5_ADDR_OF(create_srq_in, create_in, srq_context_entry); - pas = MLX5_ADDR_OF(create_srq_in, create_in, pas); - - set_srqc(srqc, in); - memcpy(pas, in->pas, pas_size); - - MLX5_SET(create_srq_in, create_in, opcode, - MLX5_CMD_OP_CREATE_SRQ); - - err = mlx5_cmd_exec(dev, create_in, inlen, create_out, - sizeof(create_out)); - kvfree(create_in); - if (!err) { - srq->srqn = MLX5_GET(create_srq_out, create_out, srqn); - srq->uid = in->uid; - } - - return err; -} - -static int destroy_srq_cmd(struct mlx5_core_dev *dev, - struct mlx5_core_srq *srq) -{ - u32 srq_in[MLX5_ST_SZ_DW(destroy_srq_in)] = {0}; - u32 srq_out[MLX5_ST_SZ_DW(destroy_srq_out)] = {0}; - - MLX5_SET(destroy_srq_in, srq_in, opcode, - MLX5_CMD_OP_DESTROY_SRQ); - MLX5_SET(destroy_srq_in, srq_in, srqn, srq->srqn); - MLX5_SET(destroy_srq_in, srq_in, uid, srq->uid); - - return mlx5_cmd_exec(dev, srq_in, sizeof(srq_in), - srq_out, sizeof(srq_out)); -} - -static int arm_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - u16 lwm, int is_srq) -{ - u32 srq_in[MLX5_ST_SZ_DW(arm_rq_in)] = {0}; - u32 srq_out[MLX5_ST_SZ_DW(arm_rq_out)] = {0}; - - MLX5_SET(arm_rq_in, srq_in, opcode, MLX5_CMD_OP_ARM_RQ); - MLX5_SET(arm_rq_in, srq_in, op_mod, MLX5_ARM_RQ_IN_OP_MOD_SRQ); - MLX5_SET(arm_rq_in, srq_in, srq_number, srq->srqn); - MLX5_SET(arm_rq_in, srq_in, lwm, lwm); - MLX5_SET(arm_rq_in, srq_in, uid, srq->uid); - - return mlx5_cmd_exec(dev, srq_in, sizeof(srq_in), - srq_out, sizeof(srq_out)); -} - -static int query_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - struct mlx5_srq_attr *out) -{ - u32 srq_in[MLX5_ST_SZ_DW(query_srq_in)] = {0}; - u32 *srq_out; - void *srqc; - int err; - - srq_out = kvzalloc(MLX5_ST_SZ_BYTES(query_srq_out), GFP_KERNEL); - if (!srq_out) - return -ENOMEM; - - MLX5_SET(query_srq_in, srq_in, opcode, - MLX5_CMD_OP_QUERY_SRQ); - MLX5_SET(query_srq_in, srq_in, srqn, srq->srqn); - err = mlx5_cmd_exec(dev, srq_in, sizeof(srq_in), - srq_out, MLX5_ST_SZ_BYTES(query_srq_out)); - if (err) - goto out; - - srqc = MLX5_ADDR_OF(query_srq_out, srq_out, srq_context_entry); - get_srqc(srqc, out); - if (MLX5_GET(srqc, srqc, state) != MLX5_SRQC_STATE_GOOD) - out->flags |= MLX5_SRQ_FLAG_ERR; -out: - kvfree(srq_out); - return err; -} - -static int create_xrc_srq_cmd(struct mlx5_core_dev *dev, - struct mlx5_core_srq *srq, - struct mlx5_srq_attr *in) -{ - u32 create_out[MLX5_ST_SZ_DW(create_xrc_srq_out)]; - void *create_in; - void *xrc_srqc; - void *pas; - int pas_size; - int inlen; - int err; - - pas_size = get_pas_size(in); - inlen = MLX5_ST_SZ_BYTES(create_xrc_srq_in) + pas_size; - create_in = kvzalloc(inlen, GFP_KERNEL); - if (!create_in) - return -ENOMEM; - - MLX5_SET(create_xrc_srq_in, create_in, uid, in->uid); - xrc_srqc = MLX5_ADDR_OF(create_xrc_srq_in, create_in, - xrc_srq_context_entry); - pas = MLX5_ADDR_OF(create_xrc_srq_in, create_in, pas); - - set_srqc(xrc_srqc, in); - MLX5_SET(xrc_srqc, xrc_srqc, user_index, in->user_index); - memcpy(pas, in->pas, pas_size); - MLX5_SET(create_xrc_srq_in, create_in, opcode, - MLX5_CMD_OP_CREATE_XRC_SRQ); - - memset(create_out, 0, sizeof(create_out)); - err = mlx5_cmd_exec(dev, create_in, inlen, create_out, - sizeof(create_out)); - if (err) - goto out; - - srq->srqn = MLX5_GET(create_xrc_srq_out, create_out, xrc_srqn); - srq->uid = in->uid; -out: - kvfree(create_in); - return err; -} - -static int destroy_xrc_srq_cmd(struct mlx5_core_dev *dev, - struct mlx5_core_srq *srq) -{ - u32 xrcsrq_in[MLX5_ST_SZ_DW(destroy_xrc_srq_in)] = {0}; - u32 xrcsrq_out[MLX5_ST_SZ_DW(destroy_xrc_srq_out)] = {0}; - - MLX5_SET(destroy_xrc_srq_in, xrcsrq_in, opcode, - MLX5_CMD_OP_DESTROY_XRC_SRQ); - MLX5_SET(destroy_xrc_srq_in, xrcsrq_in, xrc_srqn, srq->srqn); - MLX5_SET(destroy_xrc_srq_in, xrcsrq_in, uid, srq->uid); - - return mlx5_cmd_exec(dev, xrcsrq_in, sizeof(xrcsrq_in), - xrcsrq_out, sizeof(xrcsrq_out)); -} - -static int arm_xrc_srq_cmd(struct mlx5_core_dev *dev, - struct mlx5_core_srq *srq, u16 lwm) -{ - u32 xrcsrq_in[MLX5_ST_SZ_DW(arm_xrc_srq_in)] = {0}; - u32 xrcsrq_out[MLX5_ST_SZ_DW(arm_xrc_srq_out)] = {0}; - - MLX5_SET(arm_xrc_srq_in, xrcsrq_in, opcode, MLX5_CMD_OP_ARM_XRC_SRQ); - MLX5_SET(arm_xrc_srq_in, xrcsrq_in, op_mod, MLX5_ARM_XRC_SRQ_IN_OP_MOD_XRC_SRQ); - MLX5_SET(arm_xrc_srq_in, xrcsrq_in, xrc_srqn, srq->srqn); - MLX5_SET(arm_xrc_srq_in, xrcsrq_in, lwm, lwm); - MLX5_SET(arm_xrc_srq_in, xrcsrq_in, uid, srq->uid); - - return mlx5_cmd_exec(dev, xrcsrq_in, sizeof(xrcsrq_in), - xrcsrq_out, sizeof(xrcsrq_out)); -} - -static int query_xrc_srq_cmd(struct mlx5_core_dev *dev, - struct mlx5_core_srq *srq, - struct mlx5_srq_attr *out) -{ - u32 xrcsrq_in[MLX5_ST_SZ_DW(query_xrc_srq_in)]; - u32 *xrcsrq_out; - void *xrc_srqc; - int err; - - xrcsrq_out = kvzalloc(MLX5_ST_SZ_BYTES(query_xrc_srq_out), GFP_KERNEL); - if (!xrcsrq_out) - return -ENOMEM; - memset(xrcsrq_in, 0, sizeof(xrcsrq_in)); - - MLX5_SET(query_xrc_srq_in, xrcsrq_in, opcode, - MLX5_CMD_OP_QUERY_XRC_SRQ); - MLX5_SET(query_xrc_srq_in, xrcsrq_in, xrc_srqn, srq->srqn); - - err = mlx5_cmd_exec(dev, xrcsrq_in, sizeof(xrcsrq_in), xrcsrq_out, - MLX5_ST_SZ_BYTES(query_xrc_srq_out)); - if (err) - goto out; - - xrc_srqc = MLX5_ADDR_OF(query_xrc_srq_out, xrcsrq_out, - xrc_srq_context_entry); - get_srqc(xrc_srqc, out); - if (MLX5_GET(xrc_srqc, xrc_srqc, state) != MLX5_XRC_SRQC_STATE_GOOD) - out->flags |= MLX5_SRQ_FLAG_ERR; - -out: - kvfree(xrcsrq_out); - return err; -} - -static int create_rmp_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - struct mlx5_srq_attr *in) -{ - void *create_in; - void *rmpc; - void *wq; - int pas_size; - int inlen; - int err; - - pas_size = get_pas_size(in); - inlen = MLX5_ST_SZ_BYTES(create_rmp_in) + pas_size; - create_in = kvzalloc(inlen, GFP_KERNEL); - if (!create_in) - return -ENOMEM; - - rmpc = MLX5_ADDR_OF(create_rmp_in, create_in, ctx); - wq = MLX5_ADDR_OF(rmpc, rmpc, wq); - - MLX5_SET(rmpc, rmpc, state, MLX5_RMPC_STATE_RDY); - MLX5_SET(create_rmp_in, create_in, uid, in->uid); - set_wq(wq, in); - memcpy(MLX5_ADDR_OF(rmpc, rmpc, wq.pas), in->pas, pas_size); - - err = mlx5_core_create_rmp(dev, create_in, inlen, &srq->srqn); - if (!err) - srq->uid = in->uid; - - kvfree(create_in); - return err; -} - -static int destroy_rmp_cmd(struct mlx5_core_dev *dev, - struct mlx5_core_srq *srq) -{ - u32 in[MLX5_ST_SZ_DW(destroy_rmp_in)] = {}; - u32 out[MLX5_ST_SZ_DW(destroy_rmp_out)] = {}; - - MLX5_SET(destroy_rmp_in, in, opcode, MLX5_CMD_OP_DESTROY_RMP); - MLX5_SET(destroy_rmp_in, in, rmpn, srq->srqn); - MLX5_SET(destroy_rmp_in, in, uid, srq->uid); - return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); -} - -static int arm_rmp_cmd(struct mlx5_core_dev *dev, - struct mlx5_core_srq *srq, - u16 lwm) -{ - void *in; - void *rmpc; - void *wq; - void *bitmask; - int err; - - in = kvzalloc(MLX5_ST_SZ_BYTES(modify_rmp_in), GFP_KERNEL); - if (!in) - return -ENOMEM; - - rmpc = MLX5_ADDR_OF(modify_rmp_in, in, ctx); - bitmask = MLX5_ADDR_OF(modify_rmp_in, in, bitmask); - wq = MLX5_ADDR_OF(rmpc, rmpc, wq); - - MLX5_SET(modify_rmp_in, in, rmp_state, MLX5_RMPC_STATE_RDY); - MLX5_SET(modify_rmp_in, in, rmpn, srq->srqn); - MLX5_SET(modify_rmp_in, in, uid, srq->uid); - MLX5_SET(wq, wq, lwm, lwm); - MLX5_SET(rmp_bitmask, bitmask, lwm, 1); - MLX5_SET(rmpc, rmpc, state, MLX5_RMPC_STATE_RDY); - - err = mlx5_core_modify_rmp(dev, in, MLX5_ST_SZ_BYTES(modify_rmp_in)); - - kvfree(in); - return err; -} - -static int query_rmp_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - struct mlx5_srq_attr *out) -{ - u32 *rmp_out; - void *rmpc; - int err; - - rmp_out = kvzalloc(MLX5_ST_SZ_BYTES(query_rmp_out), GFP_KERNEL); - if (!rmp_out) - return -ENOMEM; - - err = mlx5_core_query_rmp(dev, srq->srqn, rmp_out); - if (err) - goto out; - - rmpc = MLX5_ADDR_OF(query_rmp_out, rmp_out, rmp_context); - get_wq(MLX5_ADDR_OF(rmpc, rmpc, wq), out); - if (MLX5_GET(rmpc, rmpc, state) != MLX5_RMPC_STATE_RDY) - out->flags |= MLX5_SRQ_FLAG_ERR; - -out: - kvfree(rmp_out); - return err; -} - -static int create_xrq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - struct mlx5_srq_attr *in) -{ - u32 create_out[MLX5_ST_SZ_DW(create_xrq_out)] = {0}; - void *create_in; - void *xrqc; - void *wq; - int pas_size; - int inlen; - int err; - - pas_size = get_pas_size(in); - inlen = MLX5_ST_SZ_BYTES(create_xrq_in) + pas_size; - create_in = kvzalloc(inlen, GFP_KERNEL); - if (!create_in) - return -ENOMEM; - - xrqc = MLX5_ADDR_OF(create_xrq_in, create_in, xrq_context); - wq = MLX5_ADDR_OF(xrqc, xrqc, wq); - - set_wq(wq, in); - memcpy(MLX5_ADDR_OF(xrqc, xrqc, wq.pas), in->pas, pas_size); - - if (in->type == IB_SRQT_TM) { - MLX5_SET(xrqc, xrqc, topology, MLX5_XRQC_TOPOLOGY_TAG_MATCHING); - if (in->flags & MLX5_SRQ_FLAG_RNDV) - MLX5_SET(xrqc, xrqc, offload, MLX5_XRQC_OFFLOAD_RNDV); - MLX5_SET(xrqc, xrqc, - tag_matching_topology_context.log_matching_list_sz, - in->tm_log_list_size); - } - MLX5_SET(xrqc, xrqc, user_index, in->user_index); - MLX5_SET(xrqc, xrqc, cqn, in->cqn); - MLX5_SET(create_xrq_in, create_in, opcode, MLX5_CMD_OP_CREATE_XRQ); - MLX5_SET(create_xrq_in, create_in, uid, in->uid); - err = mlx5_cmd_exec(dev, create_in, inlen, create_out, - sizeof(create_out)); - kvfree(create_in); - if (!err) { - srq->srqn = MLX5_GET(create_xrq_out, create_out, xrqn); - srq->uid = in->uid; - } - - return err; -} - -static int destroy_xrq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq) -{ - u32 in[MLX5_ST_SZ_DW(destroy_xrq_in)] = {0}; - u32 out[MLX5_ST_SZ_DW(destroy_xrq_out)] = {0}; - - MLX5_SET(destroy_xrq_in, in, opcode, MLX5_CMD_OP_DESTROY_XRQ); - MLX5_SET(destroy_xrq_in, in, xrqn, srq->srqn); - MLX5_SET(destroy_xrq_in, in, uid, srq->uid); - - return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); -} - -static int arm_xrq_cmd(struct mlx5_core_dev *dev, - struct mlx5_core_srq *srq, - u16 lwm) -{ - u32 out[MLX5_ST_SZ_DW(arm_rq_out)] = {0}; - u32 in[MLX5_ST_SZ_DW(arm_rq_in)] = {0}; - - MLX5_SET(arm_rq_in, in, opcode, MLX5_CMD_OP_ARM_RQ); - MLX5_SET(arm_rq_in, in, op_mod, MLX5_ARM_RQ_IN_OP_MOD_XRQ); - MLX5_SET(arm_rq_in, in, srq_number, srq->srqn); - MLX5_SET(arm_rq_in, in, lwm, lwm); - MLX5_SET(arm_rq_in, in, uid, srq->uid); - - return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); -} - -static int query_xrq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - struct mlx5_srq_attr *out) -{ - u32 in[MLX5_ST_SZ_DW(query_xrq_in)] = {0}; - u32 *xrq_out; - int outlen = MLX5_ST_SZ_BYTES(query_xrq_out); - void *xrqc; - int err; - - xrq_out = kvzalloc(outlen, GFP_KERNEL); - if (!xrq_out) - return -ENOMEM; - - MLX5_SET(query_xrq_in, in, opcode, MLX5_CMD_OP_QUERY_XRQ); - MLX5_SET(query_xrq_in, in, xrqn, srq->srqn); - - err = mlx5_cmd_exec(dev, in, sizeof(in), xrq_out, outlen); - if (err) - goto out; - - xrqc = MLX5_ADDR_OF(query_xrq_out, xrq_out, xrq_context); - get_wq(MLX5_ADDR_OF(xrqc, xrqc, wq), out); - if (MLX5_GET(xrqc, xrqc, state) != MLX5_XRQC_STATE_GOOD) - out->flags |= MLX5_SRQ_FLAG_ERR; - out->tm_next_tag = - MLX5_GET(xrqc, xrqc, - tag_matching_topology_context.append_next_index); - out->tm_hw_phase_cnt = - MLX5_GET(xrqc, xrqc, - tag_matching_topology_context.hw_phase_cnt); - out->tm_sw_phase_cnt = - MLX5_GET(xrqc, xrqc, - tag_matching_topology_context.sw_phase_cnt); - -out: - kvfree(xrq_out); - return err; -} - -static int create_srq_split(struct mlx5_core_dev *dev, - struct mlx5_core_srq *srq, - struct mlx5_srq_attr *in) -{ - if (!dev->issi) - return create_srq_cmd(dev, srq, in); - switch (srq->common.res) { - case MLX5_RES_XSRQ: - return create_xrc_srq_cmd(dev, srq, in); - case MLX5_RES_XRQ: - return create_xrq_cmd(dev, srq, in); - default: - return create_rmp_cmd(dev, srq, in); - } -} - -static int destroy_srq_split(struct mlx5_core_dev *dev, - struct mlx5_core_srq *srq) -{ - if (!dev->issi) - return destroy_srq_cmd(dev, srq); - switch (srq->common.res) { - case MLX5_RES_XSRQ: - return destroy_xrc_srq_cmd(dev, srq); - case MLX5_RES_XRQ: - return destroy_xrq_cmd(dev, srq); - default: - return destroy_rmp_cmd(dev, srq); - } -} - -int mlx5_core_create_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - struct mlx5_srq_attr *in) -{ - int err; - struct mlx5_srq_table *table = &dev->priv.srq_table; - - switch (in->type) { - case IB_SRQT_XRC: - srq->common.res = MLX5_RES_XSRQ; - break; - case IB_SRQT_TM: - srq->common.res = MLX5_RES_XRQ; - break; - default: - srq->common.res = MLX5_RES_SRQ; - } - - err = create_srq_split(dev, srq, in); - if (err) - return err; - - atomic_set(&srq->refcount, 1); - init_completion(&srq->free); - - spin_lock_irq(&table->lock); - err = radix_tree_insert(&table->tree, srq->srqn, srq); - spin_unlock_irq(&table->lock); - if (err) - goto err_destroy_srq_split; - - return 0; - -err_destroy_srq_split: - destroy_srq_split(dev, srq); - - return err; -} -EXPORT_SYMBOL(mlx5_core_create_srq); - -int mlx5_core_destroy_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq) -{ - struct mlx5_srq_table *table = &dev->priv.srq_table; - struct mlx5_core_srq *tmp; - int err; - - spin_lock_irq(&table->lock); - tmp = radix_tree_delete(&table->tree, srq->srqn); - spin_unlock_irq(&table->lock); - if (!tmp || tmp != srq) - return -EINVAL; - - err = destroy_srq_split(dev, srq); - if (err) - return err; - - if (atomic_dec_and_test(&srq->refcount)) - complete(&srq->free); - wait_for_completion(&srq->free); - - return 0; -} -EXPORT_SYMBOL(mlx5_core_destroy_srq); - -int mlx5_core_query_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - struct mlx5_srq_attr *out) -{ - if (!dev->issi) - return query_srq_cmd(dev, srq, out); - switch (srq->common.res) { - case MLX5_RES_XSRQ: - return query_xrc_srq_cmd(dev, srq, out); - case MLX5_RES_XRQ: - return query_xrq_cmd(dev, srq, out); - default: - return query_rmp_cmd(dev, srq, out); - } -} -EXPORT_SYMBOL(mlx5_core_query_srq); - -int mlx5_core_arm_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - u16 lwm, int is_srq) -{ - if (!dev->issi) - return arm_srq_cmd(dev, srq, lwm, is_srq); - switch (srq->common.res) { - case MLX5_RES_XSRQ: - return arm_xrc_srq_cmd(dev, srq, lwm); - case MLX5_RES_XRQ: - return arm_xrq_cmd(dev, srq, lwm); - default: - return arm_rmp_cmd(dev, srq, lwm); - } -} -EXPORT_SYMBOL(mlx5_core_arm_srq); - void mlx5_init_srq_table(struct mlx5_core_dev *dev) { struct mlx5_srq_table *table = &dev->priv.srq_table; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/transobj.c b/drivers/net/ethernet/mellanox/mlx5/core/transobj.c index ab482124e901..c4d4b76096dc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/transobj.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/transobj.c @@ -258,49 +258,6 @@ void mlx5_core_destroy_tis(struct mlx5_core_dev *dev, u32 tisn) } EXPORT_SYMBOL(mlx5_core_destroy_tis); -int mlx5_core_create_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen, - u32 *rmpn) -{ - u32 out[MLX5_ST_SZ_DW(create_rmp_out)] = {0}; - int err; - - MLX5_SET(create_rmp_in, in, opcode, MLX5_CMD_OP_CREATE_RMP); - err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out)); - if (!err) - *rmpn = MLX5_GET(create_rmp_out, out, rmpn); - - return err; -} - -int mlx5_core_modify_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen) -{ - u32 out[MLX5_ST_SZ_DW(modify_rmp_out)] = {0}; - - MLX5_SET(modify_rmp_in, in, opcode, MLX5_CMD_OP_MODIFY_RMP); - return mlx5_cmd_exec(dev, in, inlen, out, sizeof(out)); -} - -int mlx5_core_destroy_rmp(struct mlx5_core_dev *dev, u32 rmpn) -{ - u32 in[MLX5_ST_SZ_DW(destroy_rmp_in)] = {0}; - u32 out[MLX5_ST_SZ_DW(destroy_rmp_out)] = {0}; - - MLX5_SET(destroy_rmp_in, in, opcode, MLX5_CMD_OP_DESTROY_RMP); - MLX5_SET(destroy_rmp_in, in, rmpn, rmpn); - return mlx5_cmd_exec(dev, in, sizeof(in), out, - sizeof(out)); -} - -int mlx5_core_query_rmp(struct mlx5_core_dev *dev, u32 rmpn, u32 *out) -{ - u32 in[MLX5_ST_SZ_DW(query_rmp_in)] = {0}; - int outlen = MLX5_ST_SZ_BYTES(query_rmp_out); - - MLX5_SET(query_rmp_in, in, opcode, MLX5_CMD_OP_QUERY_RMP); - MLX5_SET(query_rmp_in, in, rmpn, rmpn); - return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen); -} - int mlx5_core_create_rqt(struct mlx5_core_dev *dev, u32 *in, int inlen, u32 *rqtn) { diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 27a481b159ed..1096da4fb368 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -904,13 +904,6 @@ struct mlx5_cmd_mailbox *mlx5_alloc_cmd_mailbox_chain(struct mlx5_core_dev *dev, gfp_t flags, int npages); void mlx5_free_cmd_mailbox_chain(struct mlx5_core_dev *dev, struct mlx5_cmd_mailbox *head); -int mlx5_core_create_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - struct mlx5_srq_attr *in); -int mlx5_core_destroy_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq); -int mlx5_core_query_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - struct mlx5_srq_attr *out); -int mlx5_core_arm_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, - u16 lwm, int is_srq); void mlx5_init_mkey_table(struct mlx5_core_dev *dev); void mlx5_cleanup_mkey_table(struct mlx5_core_dev *dev); int mlx5_core_create_mkey_cb(struct mlx5_core_dev *dev, @@ -942,7 +935,6 @@ void mlx5_unregister_debugfs(void); void mlx5_fill_page_array(struct mlx5_frag_buf *buf, __be64 *pas); void mlx5_fill_page_frag_array(struct mlx5_frag_buf *frag_buf, __be64 *pas); -struct mlx5_core_srq *mlx5_core_get_srq(struct mlx5_core_dev *dev, u32 srqn); int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn, unsigned int *irqn); int mlx5_core_attach_mcg(struct mlx5_core_dev *dev, union ib_gid *mgid, u32 qpn); diff --git a/include/linux/mlx5/srq.h b/include/linux/mlx5/srq.h index 77bc4264066d..9343306cd188 100644 --- a/include/linux/mlx5/srq.h +++ b/include/linux/mlx5/srq.h @@ -6,37 +6,6 @@ #ifndef MLX5_SRQ_H #define MLX5_SRQ_H -#include - -enum { - MLX5_SRQ_FLAG_ERR = (1 << 0), - MLX5_SRQ_FLAG_WQ_SIG = (1 << 1), - MLX5_SRQ_FLAG_RNDV = (1 << 2), -}; - -struct mlx5_srq_attr { - u32 type; - u32 flags; - u32 log_size; - u32 wqe_shift; - u32 log_page_size; - u32 wqe_cnt; - u32 srqn; - u32 xrcd; - u32 page_offset; - u32 cqn; - u32 pd; - u32 lwm; - u32 user_index; - u64 db_record; - __be64 *pas; - u32 tm_log_list_size; - u32 tm_next_tag; - u32 tm_hw_phase_cnt; - u32 tm_sw_phase_cnt; - u16 uid; -}; - struct mlx5_core_dev; void mlx5_init_srq_table(struct mlx5_core_dev *dev); diff --git a/include/linux/mlx5/transobj.h b/include/linux/mlx5/transobj.h index 39ebb699875b..a261d5528ff7 100644 --- a/include/linux/mlx5/transobj.h +++ b/include/linux/mlx5/transobj.h @@ -58,12 +58,6 @@ int mlx5_core_create_tis(struct mlx5_core_dev *dev, u32 *in, int inlen, int mlx5_core_modify_tis(struct mlx5_core_dev *dev, u32 tisn, u32 *in, int inlen); void mlx5_core_destroy_tis(struct mlx5_core_dev *dev, u32 tisn); -int mlx5_core_create_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen, - u32 *rmpn); -int mlx5_core_modify_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen); -int mlx5_core_destroy_rmp(struct mlx5_core_dev *dev, u32 rmpn); -int mlx5_core_query_rmp(struct mlx5_core_dev *dev, u32 rmpn, u32 *out); - int mlx5_core_create_rqt(struct mlx5_core_dev *dev, u32 *in, int inlen, u32 *rqtn); int mlx5_core_modify_rqt(struct mlx5_core_dev *dev, u32 rqtn, u32 *in, -- cgit v1.2.3-55-g7522 From f3da6577da67a3cd44610ca54e308c6838c92157 Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Wed, 28 Nov 2018 20:53:41 +0200 Subject: RDMA/mlx5: Initialize SRQ tables on mlx5_ib Transfer initialization and cleanup from mlx5_priv struct of mlx5_core_dev to be part of mlx5_ib_dev. This completes removal of SRQ from mlx5_core. Reviewed-by: Mark Bloch Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/mlx5/ib_rep.c | 4 ++ drivers/infiniband/hw/mlx5/main.c | 7 +++ drivers/infiniband/hw/mlx5/mlx5_ib.h | 5 +- drivers/infiniband/hw/mlx5/srq.c | 1 - drivers/infiniband/hw/mlx5/srq.h | 25 ++++++++ drivers/infiniband/hw/mlx5/srq_cmd.c | 72 ++++++++++++++++++++---- drivers/net/ethernet/mellanox/mlx5/core/Makefile | 2 +- drivers/net/ethernet/mellanox/mlx5/core/main.c | 5 -- drivers/net/ethernet/mellanox/mlx5/core/srq.c | 63 --------------------- include/linux/mlx5/driver.h | 25 -------- include/linux/mlx5/srq.h | 14 ----- 11 files changed, 101 insertions(+), 122 deletions(-) delete mode 100644 drivers/net/ethernet/mellanox/mlx5/core/srq.c delete mode 100644 include/linux/mlx5/srq.h (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/infiniband/hw/mlx5/ib_rep.c b/drivers/infiniband/hw/mlx5/ib_rep.c index 584ff2ea7810..8a682d86d634 100644 --- a/drivers/infiniband/hw/mlx5/ib_rep.c +++ b/drivers/infiniband/hw/mlx5/ib_rep.c @@ -4,6 +4,7 @@ */ #include "ib_rep.h" +#include "srq.h" static const struct mlx5_ib_profile rep_profile = { STAGE_CREATE(MLX5_IB_STAGE_INIT, @@ -21,6 +22,9 @@ static const struct mlx5_ib_profile rep_profile = { STAGE_CREATE(MLX5_IB_STAGE_ROCE, mlx5_ib_stage_rep_roce_init, mlx5_ib_stage_rep_roce_cleanup), + STAGE_CREATE(MLX5_IB_STAGE_SRQ, + mlx5_init_srq_table, + mlx5_cleanup_srq_table), STAGE_CREATE(MLX5_IB_STAGE_DEVICE_RESOURCES, mlx5_ib_stage_dev_res_init, mlx5_ib_stage_dev_res_cleanup), diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index 56472fa3e18b..96515a8c9d2c 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -60,6 +60,7 @@ #include "mlx5_ib.h" #include "ib_rep.h" #include "cmd.h" +#include "srq.h" #include #include #include @@ -6308,6 +6309,9 @@ static const struct mlx5_ib_profile pf_profile = { STAGE_CREATE(MLX5_IB_STAGE_ROCE, mlx5_ib_stage_roce_init, mlx5_ib_stage_roce_cleanup), + STAGE_CREATE(MLX5_IB_STAGE_SRQ, + mlx5_init_srq_table, + mlx5_cleanup_srq_table), STAGE_CREATE(MLX5_IB_STAGE_DEVICE_RESOURCES, mlx5_ib_stage_dev_res_init, mlx5_ib_stage_dev_res_cleanup), @@ -6365,6 +6369,9 @@ static const struct mlx5_ib_profile nic_rep_profile = { STAGE_CREATE(MLX5_IB_STAGE_ROCE, mlx5_ib_stage_rep_roce_init, mlx5_ib_stage_rep_roce_cleanup), + STAGE_CREATE(MLX5_IB_STAGE_SRQ, + mlx5_init_srq_table, + mlx5_cleanup_srq_table), STAGE_CREATE(MLX5_IB_STAGE_DEVICE_RESOURCES, mlx5_ib_stage_dev_res_init, mlx5_ib_stage_dev_res_cleanup), diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h index 332d5c4d8ab3..861b68f2e330 100644 --- a/drivers/infiniband/hw/mlx5/mlx5_ib.h +++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h @@ -41,7 +41,6 @@ #include #include #include -#include #include #include #include @@ -50,6 +49,8 @@ #include #include +#include "srq.h" + #define mlx5_ib_dbg(_dev, format, arg...) \ dev_dbg(&(_dev)->ib_dev.dev, "%s:%d:(pid %d): " format, __func__, \ __LINE__, current->pid, ##arg) @@ -774,6 +775,7 @@ enum mlx5_ib_stages { MLX5_IB_STAGE_CAPS, MLX5_IB_STAGE_NON_DEFAULT_CB, MLX5_IB_STAGE_ROCE, + MLX5_IB_STAGE_SRQ, MLX5_IB_STAGE_DEVICE_RESOURCES, MLX5_IB_STAGE_DEVICE_NOTIFIER, MLX5_IB_STAGE_ODP, @@ -942,6 +944,7 @@ struct mlx5_ib_dev { u64 sys_image_guid; struct mlx5_memic memic; u16 devx_whitelist_uid; + struct mlx5_srq_table srq_table; }; static inline struct mlx5_ib_cq *to_mibcq(struct mlx5_core_cq *mcq) diff --git a/drivers/infiniband/hw/mlx5/srq.c b/drivers/infiniband/hw/mlx5/srq.c index 2b184c7f531a..91dcd3918d96 100644 --- a/drivers/infiniband/hw/mlx5/srq.c +++ b/drivers/infiniband/hw/mlx5/srq.c @@ -5,7 +5,6 @@ #include #include -#include #include #include #include diff --git a/drivers/infiniband/hw/mlx5/srq.h b/drivers/infiniband/hw/mlx5/srq.h index 1110aeaa775e..75eb5839ae95 100644 --- a/drivers/infiniband/hw/mlx5/srq.h +++ b/drivers/infiniband/hw/mlx5/srq.h @@ -37,6 +37,28 @@ struct mlx5_srq_attr { struct mlx5_ib_dev; +struct mlx5_core_srq { + struct mlx5_core_rsc_common common; /* must be first */ + u32 srqn; + int max; + size_t max_gs; + size_t max_avail_gather; + int wqe_shift; + void (*event)(struct mlx5_core_srq *srq, enum mlx5_event e); + + atomic_t refcount; + struct completion free; + u16 uid; +}; + +struct mlx5_srq_table { + struct notifier_block nb; + /* protect radix tree + */ + spinlock_t lock; + struct radix_tree_root tree; +}; + int mlx5_cmd_create_srq(struct mlx5_ib_dev *dev, struct mlx5_core_srq *srq, struct mlx5_srq_attr *in); int mlx5_cmd_destroy_srq(struct mlx5_ib_dev *dev, struct mlx5_core_srq *srq); @@ -45,4 +67,7 @@ int mlx5_cmd_query_srq(struct mlx5_ib_dev *dev, struct mlx5_core_srq *srq, int mlx5_cmd_arm_srq(struct mlx5_ib_dev *dev, struct mlx5_core_srq *srq, u16 lwm, int is_srq); struct mlx5_core_srq *mlx5_cmd_get_srq(struct mlx5_ib_dev *dev, u32 srqn); + +int mlx5_init_srq_table(struct mlx5_ib_dev *dev); +void mlx5_cleanup_srq_table(struct mlx5_ib_dev *dev); #endif /* MLX5_IB_SRQ_H */ diff --git a/drivers/infiniband/hw/mlx5/srq_cmd.c b/drivers/infiniband/hw/mlx5/srq_cmd.c index fdb9443f49f0..6be89c6be40f 100644 --- a/drivers/infiniband/hw/mlx5/srq_cmd.c +++ b/drivers/infiniband/hw/mlx5/srq_cmd.c @@ -80,12 +80,9 @@ static void get_srqc(void *srqc, struct mlx5_srq_attr *in) struct mlx5_core_srq *mlx5_cmd_get_srq(struct mlx5_ib_dev *dev, u32 srqn) { - struct mlx5_core_dev *mdev = dev->mdev; - struct mlx5_srq_table *table; + struct mlx5_srq_table *table = &dev->srq_table; struct mlx5_core_srq *srq; - table = &mdev->priv.srq_table; - spin_lock(&table->lock); srq = radix_tree_lookup(&table->tree, srqn); @@ -576,12 +573,9 @@ static int destroy_srq_split(struct mlx5_ib_dev *dev, struct mlx5_core_srq *srq) int mlx5_cmd_create_srq(struct mlx5_ib_dev *dev, struct mlx5_core_srq *srq, struct mlx5_srq_attr *in) { - struct mlx5_core_dev *mdev = dev->mdev; - struct mlx5_srq_table *table; + struct mlx5_srq_table *table = &dev->srq_table; int err; - table = &mdev->priv.srq_table; - switch (in->type) { case IB_SRQT_XRC: srq->common.res = MLX5_RES_XSRQ; @@ -616,13 +610,10 @@ err_destroy_srq_split: int mlx5_cmd_destroy_srq(struct mlx5_ib_dev *dev, struct mlx5_core_srq *srq) { - struct mlx5_core_dev *mdev = dev->mdev; - struct mlx5_srq_table *table; + struct mlx5_srq_table *table = &dev->srq_table; struct mlx5_core_srq *tmp; int err; - table = &mdev->priv.srq_table; - spin_lock_irq(&table->lock); tmp = radix_tree_delete(&table->tree, srq->srqn); spin_unlock_irq(&table->lock); @@ -669,3 +660,60 @@ int mlx5_cmd_arm_srq(struct mlx5_ib_dev *dev, struct mlx5_core_srq *srq, return arm_rmp_cmd(dev, srq, lwm); } } + +static int srq_event_notifier(struct notifier_block *nb, + unsigned long type, void *data) +{ + struct mlx5_srq_table *table; + struct mlx5_core_srq *srq; + struct mlx5_eqe *eqe; + u32 srqn; + + if (type != MLX5_EVENT_TYPE_SRQ_CATAS_ERROR && + type != MLX5_EVENT_TYPE_SRQ_RQ_LIMIT) + return NOTIFY_DONE; + + table = container_of(nb, struct mlx5_srq_table, nb); + + eqe = data; + srqn = be32_to_cpu(eqe->data.qp_srq.qp_srq_n) & 0xffffff; + + spin_lock(&table->lock); + + srq = radix_tree_lookup(&table->tree, srqn); + if (srq) + atomic_inc(&srq->refcount); + + spin_unlock(&table->lock); + + if (!srq) + return NOTIFY_OK; + + srq->event(srq, eqe->type); + + if (atomic_dec_and_test(&srq->refcount)) + complete(&srq->free); + + return NOTIFY_OK; +} + +int mlx5_init_srq_table(struct mlx5_ib_dev *dev) +{ + struct mlx5_srq_table *table = &dev->srq_table; + + memset(table, 0, sizeof(*table)); + spin_lock_init(&table->lock); + INIT_RADIX_TREE(&table->tree, GFP_ATOMIC); + + table->nb.notifier_call = srq_event_notifier; + mlx5_notifier_register(dev->mdev, &table->nb); + + return 0; +} + +void mlx5_cleanup_srq_table(struct mlx5_ib_dev *dev) +{ + struct mlx5_srq_table *table = &dev->srq_table; + + mlx5_notifier_unregister(dev->mdev, &table->nb); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile index 26afe0779a0c..d499b3d00348 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile +++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile @@ -12,7 +12,7 @@ obj-$(CONFIG_MLX5_CORE) += mlx5_core.o # mlx5 core basic # mlx5_core-y := main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \ - health.o mcg.o cq.o srq.o alloc.o qp.o port.o mr.o pd.o \ + health.o mcg.o cq.o alloc.o qp.o port.o mr.o pd.o \ mad.o transobj.o vport.o sriov.o fs_cmd.o fs_core.o \ fs_counters.o rl.o lag.o dev.o events.o wq.o lib/gid.o \ diag/fs_tracepoint.o diag/fw_tracer.o diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 4bc27a073dc4..778995573812 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -43,7 +43,6 @@ #include #include #include -#include #include #include #include @@ -749,8 +748,6 @@ static int mlx5_init_once(struct mlx5_core_dev *dev, struct mlx5_priv *priv) mlx5_init_qp_table(dev); - mlx5_init_srq_table(dev); - mlx5_init_mkey_table(dev); mlx5_init_reserved_gids(dev); @@ -804,7 +801,6 @@ err_rl_cleanup: err_tables_cleanup: mlx5_vxlan_destroy(dev->vxlan); mlx5_cleanup_mkey_table(dev); - mlx5_cleanup_srq_table(dev); mlx5_cleanup_qp_table(dev); mlx5_cq_debugfs_cleanup(dev); err_events_cleanup: @@ -828,7 +824,6 @@ static void mlx5_cleanup_once(struct mlx5_core_dev *dev) mlx5_cleanup_clock(dev); mlx5_cleanup_reserved_gids(dev); mlx5_cleanup_mkey_table(dev); - mlx5_cleanup_srq_table(dev); mlx5_cleanup_qp_table(dev); mlx5_cq_debugfs_cleanup(dev); mlx5_events_cleanup(dev); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/srq.c b/drivers/net/ethernet/mellanox/mlx5/core/srq.c deleted file mode 100644 index 0e80ddbe2510..000000000000 --- a/drivers/net/ethernet/mellanox/mlx5/core/srq.c +++ /dev/null @@ -1,63 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB -/* - * Copyright (c) 2013-2018, Mellanox Technologies inc. All rights reserved. - */ - -#include -#include -#include - -static int srq_event_notifier(struct notifier_block *nb, - unsigned long type, void *data) -{ - struct mlx5_srq_table *table; - struct mlx5_core_srq *srq; - struct mlx5_eqe *eqe; - u32 srqn; - - if (type != MLX5_EVENT_TYPE_SRQ_CATAS_ERROR && - type != MLX5_EVENT_TYPE_SRQ_RQ_LIMIT) - return NOTIFY_DONE; - - table = container_of(nb, struct mlx5_srq_table, nb); - - eqe = data; - srqn = be32_to_cpu(eqe->data.qp_srq.qp_srq_n) & 0xffffff; - - spin_lock(&table->lock); - - srq = radix_tree_lookup(&table->tree, srqn); - if (srq) - atomic_inc(&srq->refcount); - - spin_unlock(&table->lock); - - if (!srq) - return NOTIFY_OK; - - srq->event(srq, eqe->type); - - if (atomic_dec_and_test(&srq->refcount)) - complete(&srq->free); - - return NOTIFY_OK; -} - -void mlx5_init_srq_table(struct mlx5_core_dev *dev) -{ - struct mlx5_srq_table *table = &dev->priv.srq_table; - - memset(table, 0, sizeof(*table)); - spin_lock_init(&table->lock); - INIT_RADIX_TREE(&table->tree, GFP_ATOMIC); - - table->nb.notifier_call = srq_event_notifier; - mlx5_notifier_register(dev, &table->nb); -} - -void mlx5_cleanup_srq_table(struct mlx5_core_dev *dev) -{ - struct mlx5_srq_table *table = &dev->priv.srq_table; - - mlx5_notifier_unregister(dev, &table->nb); -} diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 1096da4fb368..584d8a5df7eb 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -50,7 +50,6 @@ #include #include -#include #include #include #include @@ -393,20 +392,6 @@ struct mlx5_core_rsc_common { struct completion free; }; -struct mlx5_core_srq { - struct mlx5_core_rsc_common common; /* must be first */ - u32 srqn; - int max; - size_t max_gs; - size_t max_avail_gather; - int wqe_shift; - void (*event) (struct mlx5_core_srq *, enum mlx5_event); - - atomic_t refcount; - struct completion free; - u16 uid; -}; - struct mlx5_uars_page { void __iomem *map; bool wc; @@ -464,14 +449,6 @@ struct mlx5_qp_table { struct radix_tree_root tree; }; -struct mlx5_srq_table { - struct notifier_block nb; - /* protect radix tree - */ - spinlock_t lock; - struct radix_tree_root tree; -}; - struct mlx5_mkey_table { /* protect radix tree */ @@ -547,8 +524,6 @@ struct mlx5_priv { struct mlx5_core_health health; - struct mlx5_srq_table srq_table; - /* start: qp staff */ struct mlx5_qp_table qp_table; struct dentry *qp_debugfs; diff --git a/include/linux/mlx5/srq.h b/include/linux/mlx5/srq.h deleted file mode 100644 index 9343306cd188..000000000000 --- a/include/linux/mlx5/srq.h +++ /dev/null @@ -1,14 +0,0 @@ -/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ -/* - * Copyright (c) 2013-2018, Mellanox Technologies. All rights reserved. - */ - -#ifndef MLX5_SRQ_H -#define MLX5_SRQ_H - -struct mlx5_core_dev; - -void mlx5_init_srq_table(struct mlx5_core_dev *dev); -void mlx5_cleanup_srq_table(struct mlx5_core_dev *dev); - -#endif /* MLX5_SRQ_H */ -- cgit v1.2.3-55-g7522 From 8742c7eb3d134cced700f4ec6d7fc0c663de1e28 Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Mon, 5 Nov 2018 11:26:32 -0800 Subject: net/mlx5e: Remove trailing space of tx_pause ethtool counter name tx_pause_storm_warning_events ethtool counter name has a trailing space, remove it. Signed-off-by: Saeed Mahameed Reviewed-by: Eran Ben Elisha --- drivers/net/ethernet/mellanox/mlx5/core/en_stats.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c index 3e99d0728b2f..40b60e958cfd 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c @@ -936,7 +936,7 @@ static const struct counter_desc pport_per_prio_pfc_stats_desc[] = { }; static const struct counter_desc pport_pfc_stall_stats_desc[] = { - { "tx_pause_storm_warning_events ", PPORT_PER_PRIO_OFF(device_stall_minor_watermark_cnt) }, + { "tx_pause_storm_warning_events", PPORT_PER_PRIO_OFF(device_stall_minor_watermark_cnt) }, { "tx_pause_storm_error_events", PPORT_PER_PRIO_OFF(device_stall_critical_watermark_cnt) }, }; -- cgit v1.2.3-55-g7522 From 305438310938e9163f9366aa261336f93d2622f6 Mon Sep 17 00:00:00 2001 From: Gal Pressman Date: Wed, 7 Nov 2018 20:31:36 +0200 Subject: net/mlx5e: Cleanup unused defines Remove couple of defines that are no longer used. Signed-off-by: Gal Pressman Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 3 --- 1 file changed, 3 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 118324802926..2af43465ba2c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -147,9 +147,6 @@ struct page_pool; MLX5_UMR_MTT_ALIGNMENT)) #define MLX5E_UMR_WQEBBS \ (DIV_ROUND_UP(MLX5E_UMR_WQE_INLINE_SZ, MLX5_SEND_WQE_BB)) -#define MLX5E_ICOSQ_MAX_WQEBBS MLX5E_UMR_WQEBBS - -#define MLX5E_NUM_MAIN_GROUPS 9 #define MLX5E_MSG_LEVEL NETIF_MSG_LINK -- cgit v1.2.3-55-g7522 From 080d1b17fbc51789287cd8b2044a8cc9ec600e81 Mon Sep 17 00:00:00 2001 From: Aya Levin Date: Tue, 23 Oct 2018 10:02:08 +0300 Subject: net/mlx5e: Move modify tirs hash functionality Move modify tirs hash functionality (mlx5e_modify_tirs_hash) from en_ethtool.c to en_main.c. This allows future use of this fuctionality from en_fs_ethtool.c, while keeping current convention: en_ethtool.c doesn't have an API. There is no functional change here. Signed-off-by: Aya Levin Reviewed-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 1 + .../net/ethernet/mellanox/mlx5/core/en_ethtool.c | 25 ------------------- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 28 ++++++++++++++++++++++ 3 files changed, 29 insertions(+), 25 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 2af43465ba2c..86d60222b3cb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -799,6 +799,7 @@ int mlx5e_redirect_rqt(struct mlx5e_priv *priv, u32 rqtn, int sz, void mlx5e_build_indir_tir_ctx_hash(struct mlx5e_params *params, enum mlx5e_traffic_types tt, void *tirc, bool inner); +void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen); int mlx5e_open_locked(struct net_device *netdev); int mlx5e_close_locked(struct net_device *netdev); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c index 25c1c4f96841..2d5b00751a6d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c @@ -972,31 +972,6 @@ static int mlx5e_get_rxfh(struct net_device *netdev, u32 *indir, u8 *key, return 0; } -static void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen) -{ - void *tirc = MLX5_ADDR_OF(modify_tir_in, in, ctx); - struct mlx5_core_dev *mdev = priv->mdev; - int ctxlen = MLX5_ST_SZ_BYTES(tirc); - int tt; - - MLX5_SET(modify_tir_in, in, bitmask.hash, 1); - - for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) { - memset(tirc, 0, ctxlen); - mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, false); - mlx5_core_modify_tir(mdev, priv->indir_tir[tt].tirn, in, inlen); - } - - if (!mlx5e_tunnel_inner_ft_supported(priv->mdev)) - return; - - for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) { - memset(tirc, 0, ctxlen); - mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, true); - mlx5_core_modify_tir(mdev, priv->inner_indir_tir[tt].tirn, in, inlen); - } -} - static int mlx5e_set_rxfh(struct net_device *dev, const u32 *indir, const u8 *key, const u8 hfunc) { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 871313d6b34d..85a4633454d0 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -2735,6 +2735,34 @@ void mlx5e_build_indir_tir_ctx_hash(struct mlx5e_params *params, } } +void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen) +{ + void *tirc = MLX5_ADDR_OF(modify_tir_in, in, ctx); + struct mlx5_core_dev *mdev = priv->mdev; + int ctxlen = MLX5_ST_SZ_BYTES(tirc); + int tt; + + MLX5_SET(modify_tir_in, in, bitmask.hash, 1); + + for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) { + memset(tirc, 0, ctxlen); + mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, + false); + mlx5_core_modify_tir(mdev, priv->indir_tir[tt].tirn, in, inlen); + } + + if (!mlx5e_tunnel_inner_ft_supported(priv->mdev)) + return; + + for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) { + memset(tirc, 0, ctxlen); + mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, + true); + mlx5_core_modify_tir(mdev, priv->inner_indir_tir[tt].tirn, in, + inlen); + } +} + static int mlx5e_modify_tirs_lro(struct mlx5e_priv *priv) { struct mlx5_core_dev *mdev = priv->mdev; -- cgit v1.2.3-55-g7522 From d930ac796fdb3eed62b0053297a728ced8fd58b1 Mon Sep 17 00:00:00 2001 From: Aya Levin Date: Sun, 28 Oct 2018 16:22:57 +0200 Subject: net/mlx5e: Refactor TIR configuration function Refactor mlx5e_build_indir_tir_ctx_hash for better code re-use. TIR stands for Transport Interface Receive, which is responsible for all transport related operations on the receive side. Added a static array with TIR default configuration values. This separates configuration values from command setting, which is needed for downstream patch. Signed-off-by: Aya Levin Reviewed-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 3 +- drivers/net/ethernet/mellanox/mlx5/core/en/fs.h | 16 +++ drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 167 +++++++++------------- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 5 +- 4 files changed, 87 insertions(+), 104 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 86d60222b3cb..8539ea9a67ca 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -797,9 +797,10 @@ struct mlx5e_redirect_rqt_param { int mlx5e_redirect_rqt(struct mlx5e_priv *priv, u32 rqtn, int sz, struct mlx5e_redirect_rqt_param rrp); void mlx5e_build_indir_tir_ctx_hash(struct mlx5e_params *params, - enum mlx5e_traffic_types tt, + const struct mlx5e_tirc_config *ttconfig, void *tirc, bool inner); void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen); +struct mlx5e_tirc_config mlx5e_tirc_get_default_config(enum mlx5e_traffic_types tt); int mlx5e_open_locked(struct net_device *netdev); int mlx5e_close_locked(struct net_device *netdev); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h b/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h index 1431232c9a09..be5961ff24cc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h @@ -73,6 +73,22 @@ enum mlx5e_traffic_types { MLX5E_NUM_INDIR_TIRS = MLX5E_TT_ANY, }; +struct mlx5e_tirc_config { + u8 l3_prot_type; + u8 l4_prot_type; + u32 rx_hash_fields; +}; + +#define MLX5_HASH_IP (MLX5_HASH_FIELD_SEL_SRC_IP |\ + MLX5_HASH_FIELD_SEL_DST_IP) +#define MLX5_HASH_IP_L4PORTS (MLX5_HASH_FIELD_SEL_SRC_IP |\ + MLX5_HASH_FIELD_SEL_DST_IP |\ + MLX5_HASH_FIELD_SEL_L4_SPORT |\ + MLX5_HASH_FIELD_SEL_L4_DPORT) +#define MLX5_HASH_IP_IPSEC_SPI (MLX5_HASH_FIELD_SEL_SRC_IP |\ + MLX5_HASH_FIELD_SEL_DST_IP |\ + MLX5_HASH_FIELD_SEL_IPSEC_SPI) + enum mlx5e_tunnel_types { MLX5E_TT_IPV4_GRE, MLX5E_TT_IPV6_GRE, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 85a4633454d0..01828efe905d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -2607,6 +2607,54 @@ static void mlx5e_redirect_rqts_to_drop(struct mlx5e_priv *priv) mlx5e_redirect_rqts(priv, drop_rrp); } +static const struct mlx5e_tirc_config tirc_default_config[MLX5E_NUM_INDIR_TIRS] = { + [MLX5E_TT_IPV4_TCP] = { .l3_prot_type = MLX5_L3_PROT_TYPE_IPV4, + .l4_prot_type = MLX5_L4_PROT_TYPE_TCP, + .rx_hash_fields = MLX5_HASH_IP_L4PORTS, + }, + [MLX5E_TT_IPV6_TCP] = { .l3_prot_type = MLX5_L3_PROT_TYPE_IPV6, + .l4_prot_type = MLX5_L4_PROT_TYPE_TCP, + .rx_hash_fields = MLX5_HASH_IP_L4PORTS, + }, + [MLX5E_TT_IPV4_UDP] = { .l3_prot_type = MLX5_L3_PROT_TYPE_IPV4, + .l4_prot_type = MLX5_L4_PROT_TYPE_UDP, + .rx_hash_fields = MLX5_HASH_IP_L4PORTS, + }, + [MLX5E_TT_IPV6_UDP] = { .l3_prot_type = MLX5_L3_PROT_TYPE_IPV6, + .l4_prot_type = MLX5_L4_PROT_TYPE_UDP, + .rx_hash_fields = MLX5_HASH_IP_L4PORTS, + }, + [MLX5E_TT_IPV4_IPSEC_AH] = { .l3_prot_type = MLX5_L3_PROT_TYPE_IPV4, + .l4_prot_type = 0, + .rx_hash_fields = MLX5_HASH_IP_IPSEC_SPI, + }, + [MLX5E_TT_IPV6_IPSEC_AH] = { .l3_prot_type = MLX5_L3_PROT_TYPE_IPV6, + .l4_prot_type = 0, + .rx_hash_fields = MLX5_HASH_IP_IPSEC_SPI, + }, + [MLX5E_TT_IPV4_IPSEC_ESP] = { .l3_prot_type = MLX5_L3_PROT_TYPE_IPV4, + .l4_prot_type = 0, + .rx_hash_fields = MLX5_HASH_IP_IPSEC_SPI, + }, + [MLX5E_TT_IPV6_IPSEC_ESP] = { .l3_prot_type = MLX5_L3_PROT_TYPE_IPV6, + .l4_prot_type = 0, + .rx_hash_fields = MLX5_HASH_IP_IPSEC_SPI, + }, + [MLX5E_TT_IPV4] = { .l3_prot_type = MLX5_L3_PROT_TYPE_IPV4, + .l4_prot_type = 0, + .rx_hash_fields = MLX5_HASH_IP, + }, + [MLX5E_TT_IPV6] = { .l3_prot_type = MLX5_L3_PROT_TYPE_IPV6, + .l4_prot_type = 0, + .rx_hash_fields = MLX5_HASH_IP, + }, +}; + +struct mlx5e_tirc_config mlx5e_tirc_get_default_config(enum mlx5e_traffic_types tt) +{ + return tirc_default_config[tt]; +} + static void mlx5e_build_tir_ctx_lro(struct mlx5e_params *params, void *tirc) { if (!params->lro_en) @@ -2623,24 +2671,12 @@ static void mlx5e_build_tir_ctx_lro(struct mlx5e_params *params, void *tirc) } void mlx5e_build_indir_tir_ctx_hash(struct mlx5e_params *params, - enum mlx5e_traffic_types tt, + const struct mlx5e_tirc_config *ttconfig, void *tirc, bool inner) { void *hfso = inner ? MLX5_ADDR_OF(tirc, tirc, rx_hash_field_selector_inner) : MLX5_ADDR_OF(tirc, tirc, rx_hash_field_selector_outer); -#define MLX5_HASH_IP (MLX5_HASH_FIELD_SEL_SRC_IP |\ - MLX5_HASH_FIELD_SEL_DST_IP) - -#define MLX5_HASH_IP_L4PORTS (MLX5_HASH_FIELD_SEL_SRC_IP |\ - MLX5_HASH_FIELD_SEL_DST_IP |\ - MLX5_HASH_FIELD_SEL_L4_SPORT |\ - MLX5_HASH_FIELD_SEL_L4_DPORT) - -#define MLX5_HASH_IP_IPSEC_SPI (MLX5_HASH_FIELD_SEL_SRC_IP |\ - MLX5_HASH_FIELD_SEL_DST_IP |\ - MLX5_HASH_FIELD_SEL_IPSEC_SPI) - MLX5_SET(tirc, tirc, rx_hash_fn, mlx5e_rx_hash_fn(params->rss_hfunc)); if (params->rss_hfunc == ETH_RSS_HASH_TOP) { void *rss_key = MLX5_ADDR_OF(tirc, tirc, @@ -2651,88 +2687,12 @@ void mlx5e_build_indir_tir_ctx_hash(struct mlx5e_params *params, MLX5_SET(tirc, tirc, rx_hash_symmetric, 1); memcpy(rss_key, params->toeplitz_hash_key, len); } - - switch (tt) { - case MLX5E_TT_IPV4_TCP: - MLX5_SET(rx_hash_field_select, hfso, l3_prot_type, - MLX5_L3_PROT_TYPE_IPV4); - MLX5_SET(rx_hash_field_select, hfso, l4_prot_type, - MLX5_L4_PROT_TYPE_TCP); - MLX5_SET(rx_hash_field_select, hfso, selected_fields, - MLX5_HASH_IP_L4PORTS); - break; - - case MLX5E_TT_IPV6_TCP: - MLX5_SET(rx_hash_field_select, hfso, l3_prot_type, - MLX5_L3_PROT_TYPE_IPV6); - MLX5_SET(rx_hash_field_select, hfso, l4_prot_type, - MLX5_L4_PROT_TYPE_TCP); - MLX5_SET(rx_hash_field_select, hfso, selected_fields, - MLX5_HASH_IP_L4PORTS); - break; - - case MLX5E_TT_IPV4_UDP: - MLX5_SET(rx_hash_field_select, hfso, l3_prot_type, - MLX5_L3_PROT_TYPE_IPV4); - MLX5_SET(rx_hash_field_select, hfso, l4_prot_type, - MLX5_L4_PROT_TYPE_UDP); - MLX5_SET(rx_hash_field_select, hfso, selected_fields, - MLX5_HASH_IP_L4PORTS); - break; - - case MLX5E_TT_IPV6_UDP: - MLX5_SET(rx_hash_field_select, hfso, l3_prot_type, - MLX5_L3_PROT_TYPE_IPV6); - MLX5_SET(rx_hash_field_select, hfso, l4_prot_type, - MLX5_L4_PROT_TYPE_UDP); - MLX5_SET(rx_hash_field_select, hfso, selected_fields, - MLX5_HASH_IP_L4PORTS); - break; - - case MLX5E_TT_IPV4_IPSEC_AH: - MLX5_SET(rx_hash_field_select, hfso, l3_prot_type, - MLX5_L3_PROT_TYPE_IPV4); - MLX5_SET(rx_hash_field_select, hfso, selected_fields, - MLX5_HASH_IP_IPSEC_SPI); - break; - - case MLX5E_TT_IPV6_IPSEC_AH: - MLX5_SET(rx_hash_field_select, hfso, l3_prot_type, - MLX5_L3_PROT_TYPE_IPV6); - MLX5_SET(rx_hash_field_select, hfso, selected_fields, - MLX5_HASH_IP_IPSEC_SPI); - break; - - case MLX5E_TT_IPV4_IPSEC_ESP: - MLX5_SET(rx_hash_field_select, hfso, l3_prot_type, - MLX5_L3_PROT_TYPE_IPV4); - MLX5_SET(rx_hash_field_select, hfso, selected_fields, - MLX5_HASH_IP_IPSEC_SPI); - break; - - case MLX5E_TT_IPV6_IPSEC_ESP: - MLX5_SET(rx_hash_field_select, hfso, l3_prot_type, - MLX5_L3_PROT_TYPE_IPV6); - MLX5_SET(rx_hash_field_select, hfso, selected_fields, - MLX5_HASH_IP_IPSEC_SPI); - break; - - case MLX5E_TT_IPV4: - MLX5_SET(rx_hash_field_select, hfso, l3_prot_type, - MLX5_L3_PROT_TYPE_IPV4); - MLX5_SET(rx_hash_field_select, hfso, selected_fields, - MLX5_HASH_IP); - break; - - case MLX5E_TT_IPV6: - MLX5_SET(rx_hash_field_select, hfso, l3_prot_type, - MLX5_L3_PROT_TYPE_IPV6); - MLX5_SET(rx_hash_field_select, hfso, selected_fields, - MLX5_HASH_IP); - break; - default: - WARN_ONCE(true, "%s: bad traffic type!\n", __func__); - } + MLX5_SET(rx_hash_field_select, hfso, l3_prot_type, + ttconfig->l3_prot_type); + MLX5_SET(rx_hash_field_select, hfso, l4_prot_type, + ttconfig->l4_prot_type); + MLX5_SET(rx_hash_field_select, hfso, selected_fields, + ttconfig->rx_hash_fields); } void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen) @@ -2746,8 +2706,9 @@ void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen) for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) { memset(tirc, 0, ctxlen); - mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, - false); + mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, + &tirc_default_config[tt], + tirc, false); mlx5_core_modify_tir(mdev, priv->indir_tir[tt].tirn, in, inlen); } @@ -2756,8 +2717,9 @@ void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen) for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) { memset(tirc, 0, ctxlen); - mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, - true); + mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, + &tirc_default_config[tt], + tirc, true); mlx5_core_modify_tir(mdev, priv->inner_indir_tir[tt].tirn, in, inlen); } @@ -2816,7 +2778,8 @@ static void mlx5e_build_inner_indir_tir_ctx(struct mlx5e_priv *priv, MLX5_SET(tirc, tirc, indirect_table, priv->indir_rqt.rqtn); MLX5_SET(tirc, tirc, tunneled_offload_en, 0x1); - mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, true); + mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, + &tirc_default_config[tt], tirc, true); } static int mlx5e_set_mtu(struct mlx5_core_dev *mdev, @@ -3208,7 +3171,9 @@ static void mlx5e_build_indir_tir_ctx(struct mlx5e_priv *priv, MLX5_SET(tirc, tirc, disp_type, MLX5_TIRC_DISP_TYPE_INDIRECT); MLX5_SET(tirc, tirc, indirect_table, priv->indir_rqt.rqtn); - mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, false); + + mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, + &tirc_default_config[tt], tirc, false); } static void mlx5e_build_direct_tir_ctx(struct mlx5e_priv *priv, u32 rqtn, u32 *tirc) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index fca6f4132c91..1025afc80f42 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -360,14 +360,15 @@ static int mlx5e_hairpin_create_indirect_tirs(struct mlx5e_hairpin *hp) void *tirc; for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) { + struct mlx5e_tirc_config ttconfig = mlx5e_tirc_get_default_config(tt); + memset(in, 0, MLX5_ST_SZ_BYTES(create_tir_in)); tirc = MLX5_ADDR_OF(create_tir_in, in, ctx); MLX5_SET(tirc, tirc, transport_domain, hp->tdn); MLX5_SET(tirc, tirc, disp_type, MLX5_TIRC_DISP_TYPE_INDIRECT); MLX5_SET(tirc, tirc, indirect_table, hp->indir_rqt.rqtn); - mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, false); - + mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, &ttconfig, tirc, false); err = mlx5_core_create_tir(hp->func_mdev, in, MLX5_ST_SZ_BYTES(create_tir_in), &hp->indir_tirn[tt]); if (err) { -- cgit v1.2.3-55-g7522 From bbeb53b8b2c92fd4e5b51e8fd9ab1c4000e4d350 Mon Sep 17 00:00:00 2001 From: Aya Levin Date: Tue, 6 Nov 2018 21:05:29 +0200 Subject: net/mlx5e: Move RSS params to a dedicated struct Remove RSS params from params struct under channels, and introduce a new struct with RSS configuration params under priv struct. There is no functional change here. Signed-off-by: Aya Levin Reviewed-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 16 +++++--- .../net/ethernet/mellanox/mlx5/core/en_ethtool.c | 34 ++++++++--------- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 43 ++++++++++++---------- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 9 +++-- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 5 ++- .../net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c | 2 +- 6 files changed, 62 insertions(+), 47 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 8539ea9a67ca..330c476c1df2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -244,9 +244,6 @@ struct mlx5e_params { bool lro_en; u32 lro_wqe_sz; u8 tx_min_inline_mode; - u8 rss_hfunc; - u8 toeplitz_hash_key[40]; - u32 indirection_rqt[MLX5E_INDIR_RQT_SIZE]; bool vlan_strip_disable; bool scatter_fcs_en; bool rx_dim_enabled; @@ -651,6 +648,12 @@ enum { MLX5E_NIC_PRIO }; +struct mlx5e_rss_params { + u32 indirection_rqt[MLX5E_INDIR_RQT_SIZE]; + u8 toeplitz_hash_key[40]; + u8 hfunc; +}; + struct mlx5e_priv { /* priv data path fields - start */ struct mlx5e_txqsq *txq2sq[MLX5E_MAX_NUM_CHANNELS * MLX5E_MAX_NUM_TC]; @@ -671,6 +674,7 @@ struct mlx5e_priv { struct mlx5e_tir indir_tir[MLX5E_NUM_INDIR_TIRS]; struct mlx5e_tir inner_indir_tir[MLX5E_NUM_INDIR_TIRS]; struct mlx5e_tir direct_tir[MLX5E_MAX_NUM_CHANNELS]; + struct mlx5e_rss_params rss_params; u32 tx_rates[MLX5E_MAX_NUM_SQS]; struct mlx5e_flow_steering fs; @@ -796,7 +800,7 @@ struct mlx5e_redirect_rqt_param { int mlx5e_redirect_rqt(struct mlx5e_priv *priv, u32 rqtn, int sz, struct mlx5e_redirect_rqt_param rrp); -void mlx5e_build_indir_tir_ctx_hash(struct mlx5e_params *params, +void mlx5e_build_indir_tir_ctx_hash(struct mlx5e_rss_params *rss_params, const struct mlx5e_tirc_config *ttconfig, void *tirc, bool inner); void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen); @@ -982,11 +986,13 @@ int mlx5e_attach_netdev(struct mlx5e_priv *priv); void mlx5e_detach_netdev(struct mlx5e_priv *priv); void mlx5e_destroy_netdev(struct mlx5e_priv *priv); void mlx5e_build_nic_params(struct mlx5_core_dev *mdev, + struct mlx5e_rss_params *rss_params, struct mlx5e_params *params, u16 max_channels, u16 mtu); void mlx5e_build_rq_params(struct mlx5_core_dev *mdev, struct mlx5e_params *params); -void mlx5e_build_rss_params(struct mlx5e_params *params); +void mlx5e_build_rss_params(struct mlx5e_rss_params *rss_params, + u16 num_channels); u8 mlx5e_params_calculate_tx_min_inline(struct mlx5_core_dev *mdev); void mlx5e_rx_dim_work(struct work_struct *work); void mlx5e_tx_dim_work(struct work_struct *work); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c index 2d5b00751a6d..e868d42c83cb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c @@ -353,7 +353,7 @@ int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv, new_channels.params = priv->channels.params; new_channels.params.num_channels = count; if (!netif_is_rxfh_configured(priv->netdev)) - mlx5e_build_default_indir_rqt(new_channels.params.indirection_rqt, + mlx5e_build_default_indir_rqt(priv->rss_params.indirection_rqt, MLX5E_INDIR_RQT_SIZE, count); if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) { @@ -931,7 +931,7 @@ out: u32 mlx5e_ethtool_get_rxfh_key_size(struct mlx5e_priv *priv) { - return sizeof(priv->channels.params.toeplitz_hash_key); + return sizeof(priv->rss_params.toeplitz_hash_key); } static u32 mlx5e_get_rxfh_key_size(struct net_device *netdev) @@ -957,17 +957,18 @@ static int mlx5e_get_rxfh(struct net_device *netdev, u32 *indir, u8 *key, u8 *hfunc) { struct mlx5e_priv *priv = netdev_priv(netdev); + struct mlx5e_rss_params *rss = &priv->rss_params; if (indir) - memcpy(indir, priv->channels.params.indirection_rqt, - sizeof(priv->channels.params.indirection_rqt)); + memcpy(indir, rss->indirection_rqt, + sizeof(rss->indirection_rqt)); if (key) - memcpy(key, priv->channels.params.toeplitz_hash_key, - sizeof(priv->channels.params.toeplitz_hash_key)); + memcpy(key, rss->toeplitz_hash_key, + sizeof(rss->toeplitz_hash_key)); if (hfunc) - *hfunc = priv->channels.params.rss_hfunc; + *hfunc = rss->hfunc; return 0; } @@ -976,6 +977,7 @@ static int mlx5e_set_rxfh(struct net_device *dev, const u32 *indir, const u8 *key, const u8 hfunc) { struct mlx5e_priv *priv = netdev_priv(dev); + struct mlx5e_rss_params *rss = &priv->rss_params; int inlen = MLX5_ST_SZ_BYTES(modify_tir_in); bool hash_changed = false; void *in; @@ -991,15 +993,14 @@ static int mlx5e_set_rxfh(struct net_device *dev, const u32 *indir, mutex_lock(&priv->state_lock); - if (hfunc != ETH_RSS_HASH_NO_CHANGE && - hfunc != priv->channels.params.rss_hfunc) { - priv->channels.params.rss_hfunc = hfunc; + if (hfunc != ETH_RSS_HASH_NO_CHANGE && hfunc != rss->hfunc) { + rss->hfunc = hfunc; hash_changed = true; } if (indir) { - memcpy(priv->channels.params.indirection_rqt, indir, - sizeof(priv->channels.params.indirection_rqt)); + memcpy(rss->indirection_rqt, indir, + sizeof(rss->indirection_rqt)); if (test_bit(MLX5E_STATE_OPENED, &priv->state)) { u32 rqtn = priv->indir_rqt.rqtn; @@ -1007,7 +1008,7 @@ static int mlx5e_set_rxfh(struct net_device *dev, const u32 *indir, .is_rss = true, { .rss = { - .hfunc = priv->channels.params.rss_hfunc, + .hfunc = rss->hfunc, .channels = &priv->channels, }, }, @@ -1018,10 +1019,9 @@ static int mlx5e_set_rxfh(struct net_device *dev, const u32 *indir, } if (key) { - memcpy(priv->channels.params.toeplitz_hash_key, key, - sizeof(priv->channels.params.toeplitz_hash_key)); - hash_changed = hash_changed || - priv->channels.params.rss_hfunc == ETH_RSS_HASH_TOP; + memcpy(rss->toeplitz_hash_key, key, + sizeof(rss->toeplitz_hash_key)); + hash_changed = hash_changed || rss->hfunc == ETH_RSS_HASH_TOP; } if (hash_changed) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 01828efe905d..be782c508f3a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -2504,7 +2504,7 @@ static void mlx5e_fill_rqt_rqns(struct mlx5e_priv *priv, int sz, if (rrp.rss.hfunc == ETH_RSS_HASH_XOR) ix = mlx5e_bits_invert(i, ilog2(sz)); - ix = priv->channels.params.indirection_rqt[ix]; + ix = priv->rss_params.indirection_rqt[ix]; rqn = rrp.rss.channels->c[ix]->rq.rqn; } else { rqn = rrp.rqn; @@ -2587,7 +2587,7 @@ static void mlx5e_redirect_rqts_to_channels(struct mlx5e_priv *priv, { .rss = { .channels = chs, - .hfunc = chs->params.rss_hfunc, + .hfunc = priv->rss_params.hfunc, } }, }; @@ -2670,22 +2670,22 @@ static void mlx5e_build_tir_ctx_lro(struct mlx5e_params *params, void *tirc) MLX5_SET(tirc, tirc, lro_timeout_period_usecs, params->lro_timeout); } -void mlx5e_build_indir_tir_ctx_hash(struct mlx5e_params *params, +void mlx5e_build_indir_tir_ctx_hash(struct mlx5e_rss_params *rss_params, const struct mlx5e_tirc_config *ttconfig, void *tirc, bool inner) { void *hfso = inner ? MLX5_ADDR_OF(tirc, tirc, rx_hash_field_selector_inner) : MLX5_ADDR_OF(tirc, tirc, rx_hash_field_selector_outer); - MLX5_SET(tirc, tirc, rx_hash_fn, mlx5e_rx_hash_fn(params->rss_hfunc)); - if (params->rss_hfunc == ETH_RSS_HASH_TOP) { + MLX5_SET(tirc, tirc, rx_hash_fn, mlx5e_rx_hash_fn(rss_params->hfunc)); + if (rss_params->hfunc == ETH_RSS_HASH_TOP) { void *rss_key = MLX5_ADDR_OF(tirc, tirc, rx_hash_toeplitz_key); size_t len = MLX5_FLD_SZ_BYTES(tirc, rx_hash_toeplitz_key); MLX5_SET(tirc, tirc, rx_hash_symmetric, 1); - memcpy(rss_key, params->toeplitz_hash_key, len); + memcpy(rss_key, rss_params->toeplitz_hash_key, len); } MLX5_SET(rx_hash_field_select, hfso, l3_prot_type, ttconfig->l3_prot_type); @@ -2706,7 +2706,7 @@ void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen) for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) { memset(tirc, 0, ctxlen); - mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, + mlx5e_build_indir_tir_ctx_hash(&priv->rss_params, &tirc_default_config[tt], tirc, false); mlx5_core_modify_tir(mdev, priv->indir_tir[tt].tirn, in, inlen); @@ -2717,7 +2717,7 @@ void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen) for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) { memset(tirc, 0, ctxlen); - mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, + mlx5e_build_indir_tir_ctx_hash(&priv->rss_params, &tirc_default_config[tt], tirc, true); mlx5_core_modify_tir(mdev, priv->inner_indir_tir[tt].tirn, in, @@ -2778,7 +2778,7 @@ static void mlx5e_build_inner_indir_tir_ctx(struct mlx5e_priv *priv, MLX5_SET(tirc, tirc, indirect_table, priv->indir_rqt.rqtn); MLX5_SET(tirc, tirc, tunneled_offload_en, 0x1); - mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, + mlx5e_build_indir_tir_ctx_hash(&priv->rss_params, &tirc_default_config[tt], tirc, true); } @@ -3172,7 +3172,7 @@ static void mlx5e_build_indir_tir_ctx(struct mlx5e_priv *priv, MLX5_SET(tirc, tirc, disp_type, MLX5_TIRC_DISP_TYPE_INDIRECT); MLX5_SET(tirc, tirc, indirect_table, priv->indir_rqt.rqtn); - mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, + mlx5e_build_indir_tir_ctx_hash(&priv->rss_params, &tirc_default_config[tt], tirc, false); } @@ -4511,15 +4511,18 @@ void mlx5e_build_rq_params(struct mlx5_core_dev *mdev, mlx5e_init_rq_type_params(mdev, params); } -void mlx5e_build_rss_params(struct mlx5e_params *params) +void mlx5e_build_rss_params(struct mlx5e_rss_params *rss_params, + u16 num_channels) { - params->rss_hfunc = ETH_RSS_HASH_XOR; - netdev_rss_key_fill(params->toeplitz_hash_key, sizeof(params->toeplitz_hash_key)); - mlx5e_build_default_indir_rqt(params->indirection_rqt, - MLX5E_INDIR_RQT_SIZE, params->num_channels); + rss_params->hfunc = ETH_RSS_HASH_XOR; + netdev_rss_key_fill(rss_params->toeplitz_hash_key, + sizeof(rss_params->toeplitz_hash_key)); + mlx5e_build_default_indir_rqt(rss_params->indirection_rqt, + MLX5E_INDIR_RQT_SIZE, num_channels); } void mlx5e_build_nic_params(struct mlx5_core_dev *mdev, + struct mlx5e_rss_params *rss_params, struct mlx5e_params *params, u16 max_channels, u16 mtu) { @@ -4568,7 +4571,7 @@ void mlx5e_build_nic_params(struct mlx5_core_dev *mdev, params->tx_min_inline_mode = mlx5e_params_calculate_tx_min_inline(mdev); /* RSS */ - mlx5e_build_rss_params(params); + mlx5e_build_rss_params(rss_params, params->num_channels); } static void mlx5e_set_netdev_dev_addr(struct net_device *netdev) @@ -4741,14 +4744,16 @@ static int mlx5e_nic_init(struct mlx5_core_dev *mdev, void *ppriv) { struct mlx5e_priv *priv = netdev_priv(netdev); + struct mlx5e_rss_params *rss = &priv->rss_params; int err; err = mlx5e_netdev_init(netdev, priv, mdev, profile, ppriv); if (err) return err; - mlx5e_build_nic_params(mdev, &priv->channels.params, - mlx5e_get_netdev_max_channels(netdev), netdev->mtu); + mlx5e_build_nic_params(mdev, rss, &priv->channels.params, + mlx5e_get_netdev_max_channels(netdev), + netdev->mtu); mlx5e_timestamp_init(priv); @@ -5023,7 +5028,7 @@ int mlx5e_attach_netdev(struct mlx5e_priv *priv) if (priv->channels.params.num_channels > max_nch) { mlx5_core_warn(priv->mdev, "MLX5E: Reducing number of channels to %d\n", max_nch); priv->channels.params.num_channels = max_nch; - mlx5e_build_default_indir_rqt(priv->channels.params.indirection_rqt, + mlx5e_build_default_indir_rqt(priv->rss_params.indirection_rqt, MLX5E_INDIR_RQT_SIZE, max_nch); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index c3c657548824..a137540b84bf 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1012,7 +1012,9 @@ static const struct net_device_ops mlx5e_netdev_ops_rep = { }; static void mlx5e_build_rep_params(struct mlx5_core_dev *mdev, - struct mlx5e_params *params, u16 mtu) + struct mlx5e_params *params, + struct mlx5e_rss_params *rss_params, + u16 mtu) { u8 cq_period_mode = MLX5_CAP_GEN(mdev, cq_period_start_from_cqe) ? MLX5_CQ_PERIOD_MODE_START_FROM_CQE : @@ -1034,7 +1036,7 @@ static void mlx5e_build_rep_params(struct mlx5_core_dev *mdev, mlx5_query_min_inline(mdev, ¶ms->tx_min_inline_mode); /* RSS */ - mlx5e_build_rss_params(params); + mlx5e_build_rss_params(rss_params, params->num_channels); } static void mlx5e_build_rep_netdev(struct net_device *netdev) @@ -1087,7 +1089,8 @@ static int mlx5e_init_rep(struct mlx5_core_dev *mdev, priv->channels.params.num_channels = mlx5e_get_netdev_max_channels(netdev); - mlx5e_build_rep_params(mdev, &priv->channels.params, netdev->mtu); + mlx5e_build_rep_params(mdev, &priv->channels.params, + &priv->rss_params, netdev->mtu); mlx5e_build_rep_netdev(netdev); mlx5e_timestamp_init(priv); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 1025afc80f42..a3ff2492646d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -316,7 +316,7 @@ static void mlx5e_hairpin_fill_rqt_rqns(struct mlx5e_hairpin *hp, void *rqtc) for (i = 0; i < sz; i++) { ix = i; - if (priv->channels.params.rss_hfunc == ETH_RSS_HASH_XOR) + if (priv->rss_params.hfunc == ETH_RSS_HASH_XOR) ix = mlx5e_bits_invert(i, ilog2(sz)); ix = indirection_rqt[ix]; rqn = hp->pair->rqn[ix]; @@ -368,7 +368,8 @@ static int mlx5e_hairpin_create_indirect_tirs(struct mlx5e_hairpin *hp) MLX5_SET(tirc, tirc, transport_domain, hp->tdn); MLX5_SET(tirc, tirc, disp_type, MLX5_TIRC_DISP_TYPE_INDIRECT); MLX5_SET(tirc, tirc, indirect_table, hp->indir_rqt.rqtn); - mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, &ttconfig, tirc, false); + mlx5e_build_indir_tir_ctx_hash(&priv->rss_params, &ttconfig, tirc, false); + err = mlx5_core_create_tir(hp->func_mdev, in, MLX5_ST_SZ_BYTES(create_tir_in), &hp->indir_tirn[tt]); if (err) { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c index 11dabd62e2c7..bfc0f6581729 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c @@ -87,7 +87,7 @@ int mlx5i_init(struct mlx5_core_dev *mdev, mlx5_query_port_max_mtu(mdev, &max_mtu, 1); netdev->mtu = max_mtu; - mlx5e_build_nic_params(mdev, &priv->channels.params, + mlx5e_build_nic_params(mdev, &priv->rss_params, &priv->channels.params, mlx5e_get_netdev_max_channels(netdev), netdev->mtu); mlx5i_build_nic_params(mdev, &priv->channels.params); -- cgit v1.2.3-55-g7522 From 756c41603a189e4ae9dca77c3c3d80a926d8bedf Mon Sep 17 00:00:00 2001 From: Aya Levin Date: Tue, 23 Oct 2018 16:03:33 +0300 Subject: net/mlx5e: ethtool, Support user configuration for RX hash fields Enable user configuration of RX hash fields that are used for traffic spreading into RX queues. User can change built-in RSS (Receive Side Scaling) profiles on the following traffic types: UDP4, UDP6, TCP4 and TCP6. This configuration effects both outer and inner headers. Added support for ethtool commands: ETHTOOL_SRXFH and ETHTOOL_GRXFH. Command example respectively: $ethtool -N eth1 rx-flow-hash tcp4 sdfn $ethtool -n eth1 rx-flow-hash tcpp4 IP SA IP DA L4 bytes 0 & 1 [TCP/UDP src port] L4 bytes 2 & 3 [TCP/UDP dst port] Signed-off-by: Aya Levin Reviewed-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 1 + .../ethernet/mellanox/mlx5/core/en_fs_ethtool.c | 112 +++++++++++++++++++++ drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 27 +++-- 3 files changed, 134 insertions(+), 6 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 330c476c1df2..c6686256a92a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -650,6 +650,7 @@ enum { struct mlx5e_rss_params { u32 indirection_rqt[MLX5E_INDIR_RQT_SIZE]; + u32 rx_hash_fields[MLX5E_NUM_INDIR_TIRS]; u8 toeplitz_hash_key[40]; u8 hfunc; }; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c index c18dcebe1462..4421c10f58ae 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c @@ -771,6 +771,112 @@ void mlx5e_ethtool_init_steering(struct mlx5e_priv *priv) INIT_LIST_HEAD(&priv->fs.ethtool.rules); } +static enum mlx5e_traffic_types flow_type_to_traffic_type(u32 flow_type) +{ + switch (flow_type) { + case TCP_V4_FLOW: + return MLX5E_TT_IPV4_TCP; + case TCP_V6_FLOW: + return MLX5E_TT_IPV6_TCP; + case UDP_V4_FLOW: + return MLX5E_TT_IPV4_UDP; + case UDP_V6_FLOW: + return MLX5E_TT_IPV6_UDP; + case AH_V4_FLOW: + return MLX5E_TT_IPV4_IPSEC_AH; + case AH_V6_FLOW: + return MLX5E_TT_IPV6_IPSEC_AH; + case ESP_V4_FLOW: + return MLX5E_TT_IPV4_IPSEC_ESP; + case ESP_V6_FLOW: + return MLX5E_TT_IPV6_IPSEC_ESP; + case IPV4_FLOW: + return MLX5E_TT_IPV4; + case IPV6_FLOW: + return MLX5E_TT_IPV6; + default: + return MLX5E_NUM_INDIR_TIRS; + } +} + +static int mlx5e_set_rss_hash_opt(struct mlx5e_priv *priv, + struct ethtool_rxnfc *nfc) +{ + int inlen = MLX5_ST_SZ_BYTES(modify_tir_in); + enum mlx5e_traffic_types tt; + u8 rx_hash_field = 0; + void *in; + + tt = flow_type_to_traffic_type(nfc->flow_type); + if (tt == MLX5E_NUM_INDIR_TIRS) + return -EINVAL; + + /* RSS does not support anything other than hashing to queues + * on src IP, dest IP, TCP/UDP src port and TCP/UDP dest + * port. + */ + if (nfc->flow_type != TCP_V4_FLOW && + nfc->flow_type != TCP_V6_FLOW && + nfc->flow_type != UDP_V4_FLOW && + nfc->flow_type != UDP_V6_FLOW) + return -EOPNOTSUPP; + + if (nfc->data & ~(RXH_IP_SRC | RXH_IP_DST | + RXH_L4_B_0_1 | RXH_L4_B_2_3)) + return -EOPNOTSUPP; + + if (nfc->data & RXH_IP_SRC) + rx_hash_field |= MLX5_HASH_FIELD_SEL_SRC_IP; + if (nfc->data & RXH_IP_DST) + rx_hash_field |= MLX5_HASH_FIELD_SEL_DST_IP; + if (nfc->data & RXH_L4_B_0_1) + rx_hash_field |= MLX5_HASH_FIELD_SEL_L4_SPORT; + if (nfc->data & RXH_L4_B_2_3) + rx_hash_field |= MLX5_HASH_FIELD_SEL_L4_DPORT; + + in = kvzalloc(inlen, GFP_KERNEL); + if (!in) + return -ENOMEM; + + mutex_lock(&priv->state_lock); + + if (rx_hash_field == priv->rss_params.rx_hash_fields[tt]) + goto out; + + priv->rss_params.rx_hash_fields[tt] = rx_hash_field; + mlx5e_modify_tirs_hash(priv, in, inlen); + +out: + mutex_unlock(&priv->state_lock); + kvfree(in); + return 0; +} + +static int mlx5e_get_rss_hash_opt(struct mlx5e_priv *priv, + struct ethtool_rxnfc *nfc) +{ + enum mlx5e_traffic_types tt; + u32 hash_field = 0; + + tt = flow_type_to_traffic_type(nfc->flow_type); + if (tt == MLX5E_NUM_INDIR_TIRS) + return -EINVAL; + + hash_field = priv->rss_params.rx_hash_fields[tt]; + nfc->data = 0; + + if (hash_field & MLX5_HASH_FIELD_SEL_SRC_IP) + nfc->data |= RXH_IP_SRC; + if (hash_field & MLX5_HASH_FIELD_SEL_DST_IP) + nfc->data |= RXH_IP_DST; + if (hash_field & MLX5_HASH_FIELD_SEL_L4_SPORT) + nfc->data |= RXH_L4_B_0_1; + if (hash_field & MLX5_HASH_FIELD_SEL_L4_DPORT) + nfc->data |= RXH_L4_B_2_3; + + return 0; +} + int mlx5e_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd) { int err = 0; @@ -783,6 +889,9 @@ int mlx5e_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd) case ETHTOOL_SRXCLSRLDEL: err = mlx5e_ethtool_flow_remove(priv, cmd->fs.location); break; + case ETHTOOL_SRXFH: + err = mlx5e_set_rss_hash_opt(priv, cmd); + break; default: err = -EOPNOTSUPP; break; @@ -810,6 +919,9 @@ int mlx5e_get_rxnfc(struct net_device *dev, case ETHTOOL_GRXCLSRLALL: err = mlx5e_ethtool_get_all_flows(priv, info, rule_locs); break; + case ETHTOOL_GRXFH: + err = mlx5e_get_rss_hash_opt(priv, info); + break; default: err = -EOPNOTSUPP; break; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index be782c508f3a..544c287c6481 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -2695,20 +2695,30 @@ void mlx5e_build_indir_tir_ctx_hash(struct mlx5e_rss_params *rss_params, ttconfig->rx_hash_fields); } +static void mlx5e_update_rx_hash_fields(struct mlx5e_tirc_config *ttconfig, + enum mlx5e_traffic_types tt, + u32 rx_hash_fields) +{ + *ttconfig = tirc_default_config[tt]; + ttconfig->rx_hash_fields = rx_hash_fields; +} + void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen) { void *tirc = MLX5_ADDR_OF(modify_tir_in, in, ctx); + struct mlx5e_rss_params *rss = &priv->rss_params; struct mlx5_core_dev *mdev = priv->mdev; int ctxlen = MLX5_ST_SZ_BYTES(tirc); + struct mlx5e_tirc_config ttconfig; int tt; MLX5_SET(modify_tir_in, in, bitmask.hash, 1); for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) { memset(tirc, 0, ctxlen); - mlx5e_build_indir_tir_ctx_hash(&priv->rss_params, - &tirc_default_config[tt], - tirc, false); + mlx5e_update_rx_hash_fields(&ttconfig, tt, + rss->rx_hash_fields[tt]); + mlx5e_build_indir_tir_ctx_hash(rss, &ttconfig, tirc, false); mlx5_core_modify_tir(mdev, priv->indir_tir[tt].tirn, in, inlen); } @@ -2717,9 +2727,9 @@ void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen) for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) { memset(tirc, 0, ctxlen); - mlx5e_build_indir_tir_ctx_hash(&priv->rss_params, - &tirc_default_config[tt], - tirc, true); + mlx5e_update_rx_hash_fields(&ttconfig, tt, + rss->rx_hash_fields[tt]); + mlx5e_build_indir_tir_ctx_hash(rss, &ttconfig, tirc, true); mlx5_core_modify_tir(mdev, priv->inner_indir_tir[tt].tirn, in, inlen); } @@ -4514,11 +4524,16 @@ void mlx5e_build_rq_params(struct mlx5_core_dev *mdev, void mlx5e_build_rss_params(struct mlx5e_rss_params *rss_params, u16 num_channels) { + enum mlx5e_traffic_types tt; + rss_params->hfunc = ETH_RSS_HASH_XOR; netdev_rss_key_fill(rss_params->toeplitz_hash_key, sizeof(rss_params->toeplitz_hash_key)); mlx5e_build_default_indir_rqt(rss_params->indirection_rqt, MLX5E_INDIR_RQT_SIZE, num_channels); + for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) + rss_params->rx_hash_fields[tt] = + tirc_default_config[tt].rx_hash_fields; } void mlx5e_build_nic_params(struct mlx5_core_dev *mdev, -- cgit v1.2.3-55-g7522 From 8ff57c18e9f6b03722070a372cdcc850b13bcbc8 Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Mon, 19 Nov 2018 15:07:07 +0200 Subject: net/mlx5e: Improve ethtool private-flags code structure Refactor the code of private-flags setter. Replace consecutive calls to mlx5e_handle_pflag with a loop that uses a preset set of parameters. Signed-off-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 18 ++--- .../net/ethernet/mellanox/mlx5/core/en_ethtool.c | 76 +++++++++------------- 2 files changed, 41 insertions(+), 53 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index c6686256a92a..8760d10ad5a1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -49,6 +49,7 @@ #include #include #include +#include #include "wq.h" #include "mlx5_core.h" #include "en_stats.h" @@ -211,22 +212,23 @@ struct mlx5e_umr_wqe { extern const char mlx5e_self_tests[][ETH_GSTRING_LEN]; enum mlx5e_priv_flag { - MLX5E_PFLAG_RX_CQE_BASED_MODER = (1 << 0), - MLX5E_PFLAG_TX_CQE_BASED_MODER = (1 << 1), - MLX5E_PFLAG_RX_CQE_COMPRESS = (1 << 2), - MLX5E_PFLAG_RX_STRIDING_RQ = (1 << 3), - MLX5E_PFLAG_RX_NO_CSUM_COMPLETE = (1 << 4), + MLX5E_PFLAG_RX_CQE_BASED_MODER, + MLX5E_PFLAG_TX_CQE_BASED_MODER, + MLX5E_PFLAG_RX_CQE_COMPRESS, + MLX5E_PFLAG_RX_STRIDING_RQ, + MLX5E_PFLAG_RX_NO_CSUM_COMPLETE, + MLX5E_NUM_PFLAGS, /* Keep last */ }; #define MLX5E_SET_PFLAG(params, pflag, enable) \ do { \ if (enable) \ - (params)->pflags |= (pflag); \ + (params)->pflags |= BIT(pflag); \ else \ - (params)->pflags &= ~(pflag); \ + (params)->pflags &= ~(BIT(pflag)); \ } while (0) -#define MLX5E_GET_PFLAG(params, pflag) (!!((params)->pflags & (pflag))) +#define MLX5E_GET_PFLAG(params, pflag) (!!((params)->pflags & (BIT(pflag)))) #ifdef CONFIG_MLX5_CORE_EN_DCB #define MLX5E_MAX_BW_ALLOC 100 /* Max percentage of BW allocation */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c index e868d42c83cb..da4bb62ac528 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c @@ -135,14 +135,15 @@ void mlx5e_build_ptys2ethtool_map(void) ETHTOOL_LINK_MODE_50000baseKR2_Full_BIT); } -static const char mlx5e_priv_flags[][ETH_GSTRING_LEN] = { - "rx_cqe_moder", - "tx_cqe_moder", - "rx_cqe_compress", - "rx_striding_rq", - "rx_no_csum_complete", +typedef int (*mlx5e_pflag_handler)(struct net_device *netdev, bool enable); + +struct pflag_desc { + char name[ETH_GSTRING_LEN]; + mlx5e_pflag_handler handler; }; +static const struct pflag_desc mlx5e_priv_flags[MLX5E_NUM_PFLAGS]; + int mlx5e_ethtool_get_sset_count(struct mlx5e_priv *priv, int sset) { int i, num_stats = 0; @@ -153,7 +154,7 @@ int mlx5e_ethtool_get_sset_count(struct mlx5e_priv *priv, int sset) num_stats += mlx5e_stats_grps[i].get_num_stats(priv); return num_stats; case ETH_SS_PRIV_FLAGS: - return ARRAY_SIZE(mlx5e_priv_flags); + return MLX5E_NUM_PFLAGS; case ETH_SS_TEST: return mlx5e_self_test_num(priv); /* fallthrough */ @@ -183,8 +184,9 @@ void mlx5e_ethtool_get_strings(struct mlx5e_priv *priv, u32 stringset, u8 *data) switch (stringset) { case ETH_SS_PRIV_FLAGS: - for (i = 0; i < ARRAY_SIZE(mlx5e_priv_flags); i++) - strcpy(data + i * ETH_GSTRING_LEN, mlx5e_priv_flags[i]); + for (i = 0; i < MLX5E_NUM_PFLAGS; i++) + strcpy(data + i * ETH_GSTRING_LEN, + mlx5e_priv_flags[i].name); break; case ETH_SS_TEST: @@ -1485,8 +1487,6 @@ static int mlx5e_get_module_eeprom(struct net_device *netdev, return 0; } -typedef int (*mlx5e_pflag_handler)(struct net_device *netdev, bool enable); - static int set_pflag_cqe_based_moder(struct net_device *netdev, bool enable, bool is_rx_cq) { @@ -1649,23 +1649,30 @@ static int set_pflag_rx_no_csum_complete(struct net_device *netdev, bool enable) return 0; } +static const struct pflag_desc mlx5e_priv_flags[MLX5E_NUM_PFLAGS] = { + { "rx_cqe_moder", set_pflag_rx_cqe_based_moder }, + { "tx_cqe_moder", set_pflag_tx_cqe_based_moder }, + { "rx_cqe_compress", set_pflag_rx_cqe_compress }, + { "rx_striding_rq", set_pflag_rx_striding_rq }, + { "rx_no_csum_complete", set_pflag_rx_no_csum_complete }, +}; + static int mlx5e_handle_pflag(struct net_device *netdev, u32 wanted_flags, - enum mlx5e_priv_flag flag, - mlx5e_pflag_handler pflag_handler) + enum mlx5e_priv_flag flag) { struct mlx5e_priv *priv = netdev_priv(netdev); - bool enable = !!(wanted_flags & flag); + bool enable = !!(wanted_flags & BIT(flag)); u32 changes = wanted_flags ^ priv->channels.params.pflags; int err; - if (!(changes & flag)) + if (!(changes & BIT(flag))) return 0; - err = pflag_handler(netdev, enable); + err = mlx5e_priv_flags[flag].handler(netdev, enable); if (err) { - netdev_err(netdev, "%s private flag 0x%x failed err %d\n", - enable ? "Enable" : "Disable", flag, err); + netdev_err(netdev, "%s private flag '%s' failed err %d\n", + enable ? "Enable" : "Disable", mlx5e_priv_flags[flag].name, err); return err; } @@ -1676,38 +1683,17 @@ static int mlx5e_handle_pflag(struct net_device *netdev, static int mlx5e_set_priv_flags(struct net_device *netdev, u32 pflags) { struct mlx5e_priv *priv = netdev_priv(netdev); + enum mlx5e_priv_flag pflag; int err; mutex_lock(&priv->state_lock); - err = mlx5e_handle_pflag(netdev, pflags, - MLX5E_PFLAG_RX_CQE_BASED_MODER, - set_pflag_rx_cqe_based_moder); - if (err) - goto out; - - err = mlx5e_handle_pflag(netdev, pflags, - MLX5E_PFLAG_TX_CQE_BASED_MODER, - set_pflag_tx_cqe_based_moder); - if (err) - goto out; - - err = mlx5e_handle_pflag(netdev, pflags, - MLX5E_PFLAG_RX_CQE_COMPRESS, - set_pflag_rx_cqe_compress); - if (err) - goto out; - err = mlx5e_handle_pflag(netdev, pflags, - MLX5E_PFLAG_RX_STRIDING_RQ, - set_pflag_rx_striding_rq); - if (err) - goto out; - - err = mlx5e_handle_pflag(netdev, pflags, - MLX5E_PFLAG_RX_NO_CSUM_COMPLETE, - set_pflag_rx_no_csum_complete); + for (pflag = 0; pflag < MLX5E_NUM_PFLAGS; pflag++) { + err = mlx5e_handle_pflag(netdev, pflags, pflag); + if (err) + break; + } -out: mutex_unlock(&priv->state_lock); /* Need to fix some features.. */ -- cgit v1.2.3-55-g7522 From fe206c209314f3cebe232518db9b5ad2606e34f1 Mon Sep 17 00:00:00 2001 From: Daniel Jurgens Date: Tue, 4 Dec 2018 18:03:00 -0800 Subject: net/mlx5: When fetching CQEs return CQE instead of void pointer The function is only used to retrieve CQEs, use the proper type as the return value. Signed-off-by: Daniel Jurgens Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/wq.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/wq.h b/drivers/net/ethernet/mellanox/mlx5/core/wq.h index b1293d153a58..9bc2184a46bc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/wq.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/wq.h @@ -177,7 +177,7 @@ static inline u32 mlx5_cqwq_get_ci(struct mlx5_cqwq *wq) return mlx5_cqwq_ctr2ix(wq, wq->cc); } -static inline void *mlx5_cqwq_get_wqe(struct mlx5_cqwq *wq, u32 ix) +static inline struct mlx5_cqe64 *mlx5_cqwq_get_wqe(struct mlx5_cqwq *wq, u32 ix) { return mlx5_frag_buf_get_wqe(&wq->fbc, ix); } -- cgit v1.2.3-55-g7522 From 6254adeb1f6943a66cbed892a5f683400b9db194 Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Tue, 4 Dec 2018 18:03:01 -0800 Subject: net/mlx5: Use helper to get CQE opcode Introduce and use a helper that extracts the opcode from a CQE (completion queue entry) structure. Signed-off-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 10 +++++----- drivers/net/ethernet/mellanox/mlx5/core/en_tx.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c | 2 +- include/linux/mlx5/device.h | 5 +++++ 4 files changed, 12 insertions(+), 7 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c index 79638dcbae78..31956ddd394e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c @@ -554,9 +554,9 @@ static inline void mlx5e_poll_ico_single_cqe(struct mlx5e_cq *cq, mlx5_cqwq_pop(&cq->wq); - if (unlikely((cqe->op_own >> 4) != MLX5_CQE_REQ)) { + if (unlikely(get_cqe_opcode(cqe) != MLX5_CQE_REQ)) { netdev_WARN_ONCE(cq->channel->netdev, - "Bad OP in ICOSQ CQE: 0x%x\n", cqe->op_own); + "Bad OP in ICOSQ CQE: 0x%x\n", get_cqe_opcode(cqe)); return; } @@ -898,7 +898,7 @@ mlx5e_skb_from_cqe_linear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe, prefetchw(va); /* xdp_frame data area */ prefetch(data); - if (unlikely((cqe->op_own >> 4) != MLX5_CQE_RESP_SEND)) { + if (unlikely(get_cqe_opcode(cqe) != MLX5_CQE_RESP_SEND)) { rq->stats->wqe_err++; return NULL; } @@ -930,7 +930,7 @@ mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe, u16 byte_cnt = cqe_bcnt - headlen; struct sk_buff *skb; - if (unlikely((cqe->op_own >> 4) != MLX5_CQE_RESP_SEND)) { + if (unlikely(get_cqe_opcode(cqe) != MLX5_CQE_RESP_SEND)) { rq->stats->wqe_err++; return NULL; } @@ -1148,7 +1148,7 @@ void mlx5e_handle_rx_cqe_mpwrq(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe) wi->consumed_strides += cstrides; - if (unlikely((cqe->op_own >> 4) != MLX5_CQE_RESP_SEND)) { + if (unlikely(get_cqe_opcode(cqe) != MLX5_CQE_RESP_SEND)) { rq->stats->wqe_err++; goto mpwrq_cqe_out; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c index 6dacaeba2fbf..46b5a6914d71 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c @@ -507,7 +507,7 @@ bool mlx5e_poll_tx_cq(struct mlx5e_cq *cq, int napi_budget) wqe_counter = be16_to_cpu(cqe->wqe_counter); - if (unlikely(cqe->op_own >> 4 == MLX5_CQE_REQ_ERR)) { + if (unlikely(get_cqe_opcode(cqe) == MLX5_CQE_REQ_ERR)) { if (!test_and_set_bit(MLX5E_SQ_STATE_RECOVERING, &sq->state)) { mlx5e_dump_error_cqe(sq, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c index 8ca1d1949d93..873541ef4c1b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c @@ -334,7 +334,7 @@ static void mlx5_fpga_conn_handle_cqe(struct mlx5_fpga_conn *conn, { u8 opcode, status = 0; - opcode = cqe->op_own >> 4; + opcode = get_cqe_opcode(cqe); switch (opcode) { case MLX5_CQE_REQ_ERR: diff --git a/include/linux/mlx5/device.h b/include/linux/mlx5/device.h index f7c8bebfe472..c66867c8fc2f 100644 --- a/include/linux/mlx5/device.h +++ b/include/linux/mlx5/device.h @@ -781,6 +781,11 @@ static inline u8 mlx5_get_cqe_format(struct mlx5_cqe64 *cqe) return (cqe->op_own >> 2) & 0x3; } +static inline u8 get_cqe_opcode(struct mlx5_cqe64 *cqe) +{ + return cqe->op_own >> 4; +} + static inline u8 get_cqe_lro_tcppsh(struct mlx5_cqe64 *cqe) { return (cqe->lro_tcppsh_abort_dupack >> 6) & 1; -- cgit v1.2.3-55-g7522 From 7300375f181e06ba98183509065c6b71bead442d Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Tue, 4 Dec 2018 18:03:03 -0800 Subject: net/mlx5: Move flow counters data structures from flow steering header After the following flow counters API refactoring: ("net/mlx5: Use flow counter IDs and not the wrapping cache object") flow counters private data structures mlx5_fc_cache and mlx5_fc are redundantly exposed in fs_core.h, they have nothing to do with flow steering core and they are private to fs_counter.c, this patch moves them to where they belong and reduces their exposure in the driver. Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/fs_core.h | 23 ---------------------- .../net/ethernet/mellanox/mlx5/core/fs_counters.c | 23 ++++++++++++++++++++++ 2 files changed, 23 insertions(+), 23 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h index b51ad217da32..2dc86347af58 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h @@ -145,29 +145,6 @@ struct mlx5_flow_table { struct rhltable fgs_hash; }; -struct mlx5_fc_cache { - u64 packets; - u64 bytes; - u64 lastuse; -}; - -struct mlx5_fc { - struct list_head list; - struct llist_node addlist; - struct llist_node dellist; - - /* last{packets,bytes} members are used when calculating the delta since - * last reading - */ - u64 lastpackets; - u64 lastbytes; - - u32 id; - bool aging; - - struct mlx5_fc_cache cache ____cacheline_aligned_in_smp; -}; - struct mlx5_ft_underlay_qp { struct list_head list; u32 qpn; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c index 32accd6b041b..c6c28f56aa29 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c @@ -41,6 +41,29 @@ /* Max number of counters to query in bulk read is 32K */ #define MLX5_SW_MAX_COUNTERS_BULK BIT(15) +struct mlx5_fc_cache { + u64 packets; + u64 bytes; + u64 lastuse; +}; + +struct mlx5_fc { + struct list_head list; + struct llist_node addlist; + struct llist_node dellist; + + /* last{packets,bytes} members are used when calculating the delta since + * last reading + */ + u64 lastpackets; + u64 lastbytes; + + u32 id; + bool aging; + + struct mlx5_fc_cache cache ____cacheline_aligned_in_smp; +}; + /* locking scheme: * * It is the responsibility of the user to prevent concurrent calls or bad -- cgit v1.2.3-55-g7522 From c2fb3db22d35f01774d0547b1d8e5085df193646 Mon Sep 17 00:00:00 2001 From: Mikhael Goikhman Date: Mon, 10 Dec 2018 13:15:09 -0800 Subject: net/mlx5: Rework handling of port module events Add explicit HW defined error values. For simplicity, keep counters for all statuses starting from 0, although currently status=0 is not used. Additionally, when HW signals an unexpected cable status, it is reported now rather than ignored. And status counter is now updated on errors. Signed-off-by: Mikhael Goikhman Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_stats.c | 8 +-- drivers/net/ethernet/mellanox/mlx5/core/events.c | 83 ++++++++++++++-------- drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h | 19 +++-- 3 files changed, 65 insertions(+), 45 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c index 748d23806391..881c54c12e19 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c @@ -1087,13 +1087,13 @@ static void mlx5e_grp_per_prio_update_stats(struct mlx5e_priv *priv) } static const struct counter_desc mlx5e_pme_status_desc[] = { - { "module_unplug", 8 }, + { "module_unplug", sizeof(u64) * MLX5_MODULE_STATUS_UNPLUGGED }, }; static const struct counter_desc mlx5e_pme_error_desc[] = { - { "module_bus_stuck", 16 }, /* bus stuck (I2C or data shorted) */ - { "module_high_temp", 48 }, /* high temperature */ - { "module_bad_shorted", 56 }, /* bad or shorted cable/module */ + { "module_bus_stuck", sizeof(u64) * MLX5_MODULE_EVENT_ERROR_BUS_STUCK }, + { "module_high_temp", sizeof(u64) * MLX5_MODULE_EVENT_ERROR_HIGH_TEMPERATURE }, + { "module_bad_shorted", sizeof(u64) * MLX5_MODULE_EVENT_ERROR_BAD_CABLE }, }; #define NUM_PME_STATUS_STATS ARRAY_SIZE(mlx5e_pme_status_desc) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index e92df7020a26..587d93ec905f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -157,23 +157,43 @@ static int temp_warn(struct notifier_block *nb, unsigned long type, void *data) } /* MLX5_EVENT_TYPE_PORT_MODULE_EVENT */ -static const char *mlx5_pme_status[MLX5_MODULE_STATUS_NUM] = { - "Cable plugged", /* MLX5_MODULE_STATUS_PLUGGED = 0x1 */ - "Cable unplugged", /* MLX5_MODULE_STATUS_UNPLUGGED = 0x2 */ - "Cable error", /* MLX5_MODULE_STATUS_ERROR = 0x3 */ -}; +static const char *mlx5_pme_status_to_string(enum port_module_event_status_type status) +{ + switch (status) { + case MLX5_MODULE_STATUS_PLUGGED: + return "Cable plugged"; + case MLX5_MODULE_STATUS_UNPLUGGED: + return "Cable unplugged"; + case MLX5_MODULE_STATUS_ERROR: + return "Cable error"; + default: + return "Unknown status"; + } +} -static const char *mlx5_pme_error[MLX5_MODULE_EVENT_ERROR_NUM] = { - "Power budget exceeded", - "Long Range for non MLNX cable", - "Bus stuck(I2C or data shorted)", - "No EEPROM/retry timeout", - "Enforce part number list", - "Unknown identifier", - "High Temperature", - "Bad or shorted cable/module", - "Unknown status", -}; +static const char *mlx5_pme_error_to_string(enum port_module_event_error_type error) +{ + switch (error) { + case MLX5_MODULE_EVENT_ERROR_POWER_BUDGET_EXCEEDED: + return "Power budget exceeded"; + case MLX5_MODULE_EVENT_ERROR_LONG_RANGE_FOR_NON_MLNX: + return "Long Range for non MLNX cable"; + case MLX5_MODULE_EVENT_ERROR_BUS_STUCK: + return "Bus stuck (I2C or data shorted)"; + case MLX5_MODULE_EVENT_ERROR_NO_EEPROM_RETRY_TIMEOUT: + return "No EEPROM/retry timeout"; + case MLX5_MODULE_EVENT_ERROR_ENFORCE_PART_NUMBER_LIST: + return "Enforce part number list"; + case MLX5_MODULE_EVENT_ERROR_UNKNOWN_IDENTIFIER: + return "Unknown identifier"; + case MLX5_MODULE_EVENT_ERROR_HIGH_TEMPERATURE: + return "High Temperature"; + case MLX5_MODULE_EVENT_ERROR_BAD_CABLE: + return "Bad or shorted cable/module"; + default: + return "Unknown error"; + } +} /* type == MLX5_EVENT_TYPE_PORT_MODULE_EVENT */ static int port_module(struct notifier_block *nb, unsigned long type, void *data) @@ -185,6 +205,7 @@ static int port_module(struct notifier_block *nb, unsigned long type, void *data enum port_module_event_status_type module_status; enum port_module_event_error_type error_type; struct mlx5_eqe_port_module *module_event_eqe; + const char *status_str, *error_str; u8 module_num; module_event_eqe = &eqe->data.port_module; @@ -193,28 +214,28 @@ static int port_module(struct notifier_block *nb, unsigned long type, void *data PORT_MODULE_EVENT_MODULE_STATUS_MASK; error_type = module_event_eqe->error_type & PORT_MODULE_EVENT_ERROR_TYPE_MASK; - if (module_status < MLX5_MODULE_STATUS_ERROR) { - events->pme_stats.status_counters[module_status - 1]++; - } else if (module_status == MLX5_MODULE_STATUS_ERROR) { - if (error_type >= MLX5_MODULE_EVENT_ERROR_UNKNOWN) - /* Unknown error type */ - error_type = MLX5_MODULE_EVENT_ERROR_UNKNOWN; - events->pme_stats.error_counters[error_type]++; + + if (module_status < MLX5_MODULE_STATUS_NUM) + events->pme_stats.status_counters[module_status]++; + status_str = mlx5_pme_status_to_string(module_status); + + if (module_status == MLX5_MODULE_STATUS_ERROR) { + if (error_type < MLX5_MODULE_EVENT_ERROR_NUM) + events->pme_stats.error_counters[error_type]++; + error_str = mlx5_pme_error_to_string(error_type); } if (!printk_ratelimit()) return NOTIFY_OK; - if (module_status < MLX5_MODULE_STATUS_ERROR) + if (module_status == MLX5_MODULE_STATUS_ERROR) + mlx5_core_err(events->dev, + "Port module event[error]: module %u, %s, %s\n", + module_num, status_str, error_str); + else mlx5_core_info(events->dev, "Port module event: module %u, %s\n", - module_num, mlx5_pme_status[module_status - 1]); - - else if (module_status == MLX5_MODULE_STATUS_ERROR) - mlx5_core_info(events->dev, - "Port module event[error]: module %u, %s, %s\n", - module_num, mlx5_pme_status[module_status - 1], - mlx5_pme_error[error_type]); + module_num, status_str); return NOTIFY_OK; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h index 4d78a459676e..af19fa61e9ef 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h @@ -51,19 +51,18 @@ enum port_module_event_status_type { MLX5_MODULE_STATUS_PLUGGED = 0x1, MLX5_MODULE_STATUS_UNPLUGGED = 0x2, MLX5_MODULE_STATUS_ERROR = 0x3, - MLX5_MODULE_STATUS_NUM = 0x3, + MLX5_MODULE_STATUS_NUM, }; enum port_module_event_error_type { - MLX5_MODULE_EVENT_ERROR_POWER_BUDGET_EXCEEDED, - MLX5_MODULE_EVENT_ERROR_LONG_RANGE_FOR_NON_MLNX_CABLE_MODULE, - MLX5_MODULE_EVENT_ERROR_BUS_STUCK, - MLX5_MODULE_EVENT_ERROR_NO_EEPROM_RETRY_TIMEOUT, - MLX5_MODULE_EVENT_ERROR_ENFORCE_PART_NUMBER_LIST, - MLX5_MODULE_EVENT_ERROR_UNKNOWN_IDENTIFIER, - MLX5_MODULE_EVENT_ERROR_HIGH_TEMPERATURE, - MLX5_MODULE_EVENT_ERROR_BAD_CABLE, - MLX5_MODULE_EVENT_ERROR_UNKNOWN, + MLX5_MODULE_EVENT_ERROR_POWER_BUDGET_EXCEEDED = 0x0, + MLX5_MODULE_EVENT_ERROR_LONG_RANGE_FOR_NON_MLNX = 0x1, + MLX5_MODULE_EVENT_ERROR_BUS_STUCK = 0x2, + MLX5_MODULE_EVENT_ERROR_NO_EEPROM_RETRY_TIMEOUT = 0x3, + MLX5_MODULE_EVENT_ERROR_ENFORCE_PART_NUMBER_LIST = 0x4, + MLX5_MODULE_EVENT_ERROR_UNKNOWN_IDENTIFIER = 0x5, + MLX5_MODULE_EVENT_ERROR_HIGH_TEMPERATURE = 0x6, + MLX5_MODULE_EVENT_ERROR_BAD_CABLE = 0x7, MLX5_MODULE_EVENT_ERROR_NUM, }; -- cgit v1.2.3-55-g7522 From 37a12aae06fda00f36eadb168d5d7ba0be3a8c86 Mon Sep 17 00:00:00 2001 From: Mikhael Goikhman Date: Mon, 10 Dec 2018 13:15:10 -0800 Subject: net/mlx5: Add support for PCIe power slot exceeded error in PME Support a new hardware error type in port module events: - error_type=0xc (PCIe system power slot exceeded) Signed-off-by: Mikhael Goikhman Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/events.c | 2 ++ drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h | 1 + 2 files changed, 3 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index 587d93ec905f..ba9249bcfcba 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -190,6 +190,8 @@ static const char *mlx5_pme_error_to_string(enum port_module_event_error_type er return "High Temperature"; case MLX5_MODULE_EVENT_ERROR_BAD_CABLE: return "Bad or shorted cable/module"; + case MLX5_MODULE_EVENT_ERROR_PCIE_POWER_SLOT_EXCEEDED: + return "One or more network ports have been powered down due to insufficient/unadvertised power on the PCIe slot"; default: return "Unknown error"; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h index af19fa61e9ef..979970f8fa0f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h @@ -63,6 +63,7 @@ enum port_module_event_error_type { MLX5_MODULE_EVENT_ERROR_UNKNOWN_IDENTIFIER = 0x5, MLX5_MODULE_EVENT_ERROR_HIGH_TEMPERATURE = 0x6, MLX5_MODULE_EVENT_ERROR_BAD_CABLE = 0x7, + MLX5_MODULE_EVENT_ERROR_PCIE_POWER_SLOT_EXCEEDED = 0xc, MLX5_MODULE_EVENT_ERROR_NUM, }; -- cgit v1.2.3-55-g7522 From 8d6b57e6445989e9f34077ef1aac80b621029b08 Mon Sep 17 00:00:00 2001 From: Mikhael Goikhman Date: Mon, 10 Dec 2018 13:15:11 -0800 Subject: net/mlx5: Add support for plugged-disabled cable status in PME Support a new hardware module status in port module events: - module_status=0x4 (Cable plugged, but disabled) Signed-off-by: Mikhael Goikhman Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/events.c | 2 ++ drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h | 1 + 2 files changed, 3 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index ba9249bcfcba..900fdd235ba0 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -166,6 +166,8 @@ static const char *mlx5_pme_status_to_string(enum port_module_event_status_type return "Cable unplugged"; case MLX5_MODULE_STATUS_ERROR: return "Cable error"; + case MLX5_MODULE_STATUS_DISABLED: + return "Cable disabled"; default: return "Unknown status"; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h index 979970f8fa0f..397a2847867a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h @@ -51,6 +51,7 @@ enum port_module_event_status_type { MLX5_MODULE_STATUS_PLUGGED = 0x1, MLX5_MODULE_STATUS_UNPLUGGED = 0x2, MLX5_MODULE_STATUS_ERROR = 0x3, + MLX5_MODULE_STATUS_DISABLED = 0x4, MLX5_MODULE_STATUS_NUM, }; -- cgit v1.2.3-55-g7522 From fd4572b3ff3ff57ca7fa612f9ea42b90afdd8bff Mon Sep 17 00:00:00 2001 From: Eyal Davidovich Date: Mon, 10 Dec 2018 13:15:12 -0800 Subject: net/mlx5: Add monitor commands layout and event data Will be used in downstream patch to monitor counter changes by the HCA and report it to the driver by an event. The driver will update its counters cached data accordingly. Signed-off-by: Eyal Davidovich Reviewed-by: Eran Ben Elisha Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/cmd.c | 4 ++ drivers/net/ethernet/mellanox/mlx5/core/eq.c | 3 + drivers/net/ethernet/mellanox/mlx5/core/events.c | 2 + include/linux/mlx5/device.h | 1 + include/linux/mlx5/mlx5_ifc.h | 87 +++++++++++++++++++++++- 5 files changed, 96 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c index 8ab636d59edb..d3125cdf69db 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c @@ -373,6 +373,8 @@ static int mlx5_internal_err_ret_value(struct mlx5_core_dev *dev, u16 op, case MLX5_CMD_OP_QUERY_VPORT_COUNTER: case MLX5_CMD_OP_ALLOC_Q_COUNTER: case MLX5_CMD_OP_QUERY_Q_COUNTER: + case MLX5_CMD_OP_SET_MONITOR_COUNTER: + case MLX5_CMD_OP_ARM_MONITOR_COUNTER: case MLX5_CMD_OP_SET_PP_RATE_LIMIT: case MLX5_CMD_OP_QUERY_RATE_LIMIT: case MLX5_CMD_OP_CREATE_SCHEDULING_ELEMENT: @@ -522,6 +524,8 @@ const char *mlx5_command_str(int command) MLX5_COMMAND_STR_CASE(ALLOC_Q_COUNTER); MLX5_COMMAND_STR_CASE(DEALLOC_Q_COUNTER); MLX5_COMMAND_STR_CASE(QUERY_Q_COUNTER); + MLX5_COMMAND_STR_CASE(SET_MONITOR_COUNTER); + MLX5_COMMAND_STR_CASE(ARM_MONITOR_COUNTER); MLX5_COMMAND_STR_CASE(SET_PP_RATE_LIMIT); MLX5_COMMAND_STR_CASE(QUERY_RATE_LIMIT); MLX5_COMMAND_STR_CASE(CREATE_SCHEDULING_ELEMENT); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index 4aa39a1fe23f..ee04aab65a9f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -527,6 +527,9 @@ static u64 gather_async_events_mask(struct mlx5_core_dev *dev) if (MLX5_CAP_MCAM_REG(dev, tracer_registers)) async_event_mask |= (1ull << MLX5_EVENT_TYPE_DEVICE_TRACER); + if (MLX5_CAP_GEN(dev, max_num_of_monitor_counters)) + async_event_mask |= (1ull << MLX5_EVENT_TYPE_MONITOR_COUNTER); + return async_event_mask; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/events.c b/drivers/net/ethernet/mellanox/mlx5/core/events.c index 900fdd235ba0..fbc42b7252a9 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/events.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/events.c @@ -117,6 +117,8 @@ static const char *eqe_type_str(u8 type) return "MLX5_EVENT_TYPE_FPGA_QP_ERROR"; case MLX5_EVENT_TYPE_GENERAL_EVENT: return "MLX5_EVENT_TYPE_GENERAL_EVENT"; + case MLX5_EVENT_TYPE_MONITOR_COUNTER: + return "MLX5_EVENT_TYPE_MONITOR_COUNTER"; case MLX5_EVENT_TYPE_DEVICE_TRACER: return "MLX5_EVENT_TYPE_DEVICE_TRACER"; default: diff --git a/include/linux/mlx5/device.h b/include/linux/mlx5/device.h index c66867c8fc2f..4674b9e99f45 100644 --- a/include/linux/mlx5/device.h +++ b/include/linux/mlx5/device.h @@ -330,6 +330,7 @@ enum mlx5_event { MLX5_EVENT_TYPE_TEMP_WARN_EVENT = 0x17, MLX5_EVENT_TYPE_REMOTE_CONFIG = 0x19, MLX5_EVENT_TYPE_GENERAL_EVENT = 0x22, + MLX5_EVENT_TYPE_MONITOR_COUNTER = 0x24, MLX5_EVENT_TYPE_PPS_EVENT = 0x25, MLX5_EVENT_TYPE_DB_BF_CONGESTION = 0x1a, diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index 91d6e85e3cef..9f7cc26bfb3b 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -161,6 +161,8 @@ enum { MLX5_CMD_OP_ALLOC_Q_COUNTER = 0x771, MLX5_CMD_OP_DEALLOC_Q_COUNTER = 0x772, MLX5_CMD_OP_QUERY_Q_COUNTER = 0x773, + MLX5_CMD_OP_SET_MONITOR_COUNTER = 0x774, + MLX5_CMD_OP_ARM_MONITOR_COUNTER = 0x775, MLX5_CMD_OP_SET_PP_RATE_LIMIT = 0x780, MLX5_CMD_OP_QUERY_RATE_LIMIT = 0x781, MLX5_CMD_OP_CREATE_SCHEDULING_ELEMENT = 0x782, @@ -1200,7 +1202,13 @@ struct mlx5_ifc_cmd_hca_cap_bits { u8 sw_owner_id[0x1]; u8 reserved_at_61f[0x1]; - u8 reserved_at_620[0x80]; + u8 max_num_of_monitor_counters[0x10]; + u8 num_ppcnt_monitor_counters[0x10]; + + u8 reserved_at_640[0x10]; + u8 num_q_monitor_counters[0x10]; + + u8 reserved_at_660[0x40]; u8 uctx_cap[0x20]; @@ -3808,6 +3816,83 @@ enum { MLX5_VPORT_STATE_OP_MOD_ESW_VPORT = 0x1, }; +struct mlx5_ifc_arm_monitor_counter_in_bits { + u8 opcode[0x10]; + u8 uid[0x10]; + + u8 reserved_at_20[0x10]; + u8 op_mod[0x10]; + + u8 reserved_at_40[0x20]; + + u8 reserved_at_60[0x20]; +}; + +struct mlx5_ifc_arm_monitor_counter_out_bits { + u8 status[0x8]; + u8 reserved_at_8[0x18]; + + u8 syndrome[0x20]; + + u8 reserved_at_40[0x40]; +}; + +enum { + MLX5_QUERY_MONITOR_CNT_TYPE_PPCNT = 0x0, + MLX5_QUERY_MONITOR_CNT_TYPE_Q_COUNTER = 0x1, +}; + +enum mlx5_monitor_counter_ppcnt { + MLX5_QUERY_MONITOR_PPCNT_IN_RANGE_LENGTH_ERRORS = 0X0, + MLX5_QUERY_MONITOR_PPCNT_OUT_OF_RANGE_LENGTH_FIELD = 0X1, + MLX5_QUERY_MONITOR_PPCNT_FRAME_TOO_LONG_ERRORS = 0X2, + MLX5_QUERY_MONITOR_PPCNT_FRAME_CHECK_SEQUENCE_ERRORS = 0X3, + MLX5_QUERY_MONITOR_PPCNT_ALIGNMENT_ERRORS = 0X4, + MLX5_QUERY_MONITOR_PPCNT_IF_OUT_DISCARDS = 0X5, +}; + +enum { + MLX5_QUERY_MONITOR_Q_COUNTER_RX_OUT_OF_BUFFER = 0X4, +}; + +struct mlx5_ifc_monitor_counter_output_bits { + u8 reserved_at_0[0x4]; + u8 type[0x4]; + u8 reserved_at_8[0x8]; + u8 counter[0x10]; + + u8 counter_group_id[0x20]; +}; + +#define MLX5_CMD_SET_MONITOR_NUM_PPCNT_COUNTER_SET1 (6) +#define MLX5_CMD_SET_MONITOR_NUM_Q_COUNTERS_SET1 (1) +#define MLX5_CMD_SET_MONITOR_NUM_COUNTER (MLX5_CMD_SET_MONITOR_NUM_PPCNT_COUNTER_SET1 +\ + MLX5_CMD_SET_MONITOR_NUM_Q_COUNTERS_SET1) + +struct mlx5_ifc_set_monitor_counter_in_bits { + u8 opcode[0x10]; + u8 uid[0x10]; + + u8 reserved_at_20[0x10]; + u8 op_mod[0x10]; + + u8 reserved_at_40[0x10]; + u8 num_of_counters[0x10]; + + u8 reserved_at_60[0x20]; + + struct mlx5_ifc_monitor_counter_output_bits monitor_counter[MLX5_CMD_SET_MONITOR_NUM_COUNTER]; +}; + +struct mlx5_ifc_set_monitor_counter_out_bits { + u8 status[0x8]; + u8 reserved_at_8[0x18]; + + u8 syndrome[0x20]; + + u8 reserved_at_40[0x40]; +}; + struct mlx5_ifc_query_vport_state_in_bits { u8 opcode[0x10]; u8 reserved_at_10[0x10]; -- cgit v1.2.3-55-g7522 From 5886a96ad19dacebe6c4f7f8c001d489b06125dc Mon Sep 17 00:00:00 2001 From: Oz Shlomo Date: Mon, 10 Dec 2018 13:15:13 -0800 Subject: net/mlx5: Revise gre and nvgre key formats GRE RFC defines a 32 bit key field. NVGRE RFC splits the 32 bit key field to 24 bit VSID (gre_key_h) and 8 bit flow entropy (gre_key_l). Define the two key parsing alternatives in a union, thus enabling both access methods. Signed-off-by: Oz Shlomo Reviewed-by: Eli Britstein Signed-off-by: Saeed Mahameed --- drivers/infiniband/hw/mlx5/main.c | 4 ++-- .../net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c | 8 ++++---- include/linux/mlx5/mlx5_ifc.h | 13 +++++++++++-- 3 files changed, 17 insertions(+), 8 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index 96515a8c9d2c..2560996fce79 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -2680,11 +2680,11 @@ static int parse_flow_attr(struct mlx5_core_dev *mdev, u32 *match_c, ntohs(ib_spec->gre.val.protocol)); memcpy(MLX5_ADDR_OF(fte_match_set_misc, misc_params_c, - gre_key_h), + gre_key.nvgre.hi), &ib_spec->gre.mask.key, sizeof(ib_spec->gre.mask.key)); memcpy(MLX5_ADDR_OF(fte_match_set_misc, misc_params_v, - gre_key_h), + gre_key.nvgre.hi), &ib_spec->gre.val.key, sizeof(ib_spec->gre.val.key)); break; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c b/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c index 0f11fff32a9b..424457ff9759 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c @@ -161,10 +161,10 @@ static void print_misc_parameters_hdrs(struct trace_seq *p, PRINT_MASKED_VAL(name, p, format); \ } DECLARE_MASK_VAL(u64, gre_key) = { - .m = MLX5_GET(fte_match_set_misc, mask, gre_key_h) << 8 | - MLX5_GET(fte_match_set_misc, mask, gre_key_l), - .v = MLX5_GET(fte_match_set_misc, value, gre_key_h) << 8 | - MLX5_GET(fte_match_set_misc, value, gre_key_l)}; + .m = MLX5_GET(fte_match_set_misc, mask, gre_key.nvgre.hi) << 8 | + MLX5_GET(fte_match_set_misc, mask, gre_key.nvgre.lo), + .v = MLX5_GET(fte_match_set_misc, value, gre_key.nvgre.hi) << 8 | + MLX5_GET(fte_match_set_misc, value, gre_key.nvgre.lo)}; PRINT_MASKED_VAL(gre_key, p, "%llu"); PRINT_MASKED_VAL_MISC(u32, source_sqn, source_sqn, p, "%u"); diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index 9f7cc26bfb3b..688a549e74f1 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -423,6 +423,16 @@ struct mlx5_ifc_fte_match_set_lyr_2_4_bits { union mlx5_ifc_ipv6_layout_ipv4_layout_auto_bits dst_ipv4_dst_ipv6; }; +struct mlx5_ifc_nvgre_key_bits { + u8 hi[0x18]; + u8 lo[0x8]; +}; + +union mlx5_ifc_gre_key_bits { + struct mlx5_ifc_nvgre_key_bits nvgre; + u8 key[0x20]; +}; + struct mlx5_ifc_fte_match_set_misc_bits { u8 reserved_at_0[0x8]; u8 source_sqn[0x18]; @@ -444,8 +454,7 @@ struct mlx5_ifc_fte_match_set_misc_bits { u8 reserved_at_64[0xc]; u8 gre_protocol[0x10]; - u8 gre_key_h[0x18]; - u8 gre_key_l[0x8]; + union mlx5_ifc_gre_key_bits gre_key; u8 vxlan_vni[0x18]; u8 reserved_at_b8[0x8]; -- cgit v1.2.3-55-g7522 From aa39c2c0e44d16b5804f8fb6b5350cdf4e33b4c3 Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Mon, 10 Dec 2018 13:15:15 -0800 Subject: net/mlx5: E-Switch, Change vhca id valid bool field to bit flag Change the driver flow destination struct to use bit flags with the vhca id valid being the 1st one. The flags field is more extendable and will be used in downstream patch. Signed-off-by: Eli Britstein Reviewed-by: Or Gerlitz Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 8 +++++--- drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c | 3 ++- include/linux/mlx5/fs.h | 6 +++++- 3 files changed, 12 insertions(+), 5 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 9eac137790f5..4d7b65df32ef 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -125,8 +125,9 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw, dest[i].vport.num = attr->out_rep[j]->vport; dest[i].vport.vhca_id = MLX5_CAP_GEN(attr->out_mdev[j], vhca_id); - dest[i].vport.vhca_id_valid = - !!MLX5_CAP_ESW(esw->dev, merged_eswitch); + if (MLX5_CAP_ESW(esw->dev, merged_eswitch)) + dest[i].vport.flags |= + MLX5_FLOW_DEST_VPORT_VHCA_ID; i++; } } @@ -220,7 +221,8 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw, dest[i].vport.num = attr->out_rep[i]->vport; dest[i].vport.vhca_id = MLX5_CAP_GEN(attr->out_mdev[i], vhca_id); - dest[i].vport.vhca_id_valid = !!MLX5_CAP_ESW(esw->dev, merged_eswitch); + if (MLX5_CAP_ESW(esw->dev, merged_eswitch)) + dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_VHCA_ID; } dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE; dest[i].ft = fwd_fdb, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c index 08a891f9aade..dda63dedaa49 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c @@ -387,7 +387,8 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev, id = dst->dest_attr.vport.num; MLX5_SET(dest_format_struct, in_dests, destination_eswitch_owner_vhca_id_valid, - dst->dest_attr.vport.vhca_id_valid); + !!(dst->dest_attr.vport.flags & + MLX5_FLOW_DEST_VPORT_VHCA_ID)); MLX5_SET(dest_format_struct, in_dests, destination_eswitch_owner_vhca_id, dst->dest_attr.vport.vhca_id); diff --git a/include/linux/mlx5/fs.h b/include/linux/mlx5/fs.h index 5660f07d3be0..25ffd8018b72 100644 --- a/include/linux/mlx5/fs.h +++ b/include/linux/mlx5/fs.h @@ -86,6 +86,10 @@ struct mlx5_flow_spec { u32 match_value[MLX5_ST_SZ_DW(fte_match_param)]; }; +enum { + MLX5_FLOW_DEST_VPORT_VHCA_ID = BIT(0), +}; + struct mlx5_flow_destination { enum mlx5_flow_destination_type type; union { @@ -96,7 +100,7 @@ struct mlx5_flow_destination { struct { u16 num; u16 vhca_id; - bool vhca_id_valid; + u8 flags; } vport; }; }; -- cgit v1.2.3-55-g7522 From a2c6162b12f15fbbbe38d0eb3a38186bcfc79c0f Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Mon, 10 Dec 2018 13:15:16 -0800 Subject: net/mlx5: Support extended destination format in flow steering command Update the flow steering command formatting according to the extended destination API. Note that the FW dictates that multi destination FTEs that involve at least one encap must use the extended destination format, while single destination ones must use the legacy format. Using extended destination format requires FW support. Check for its capabilities and return error if not supported. Signed-off-by: Eli Britstein Reviewed-by: Or Gerlitz Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c | 80 +++++++++++++++++++++--- include/linux/mlx5/fs.h | 2 + 2 files changed, 75 insertions(+), 7 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c index dda63dedaa49..c44ccb67c4a3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c @@ -308,22 +308,68 @@ static int mlx5_cmd_destroy_flow_group(struct mlx5_core_dev *dev, return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); } +static int mlx5_set_extended_dest(struct mlx5_core_dev *dev, + struct fs_fte *fte, bool *extended_dest) +{ + int fw_log_max_fdb_encap_uplink = + MLX5_CAP_ESW(dev, log_max_fdb_encap_uplink); + int num_fwd_destinations = 0; + struct mlx5_flow_rule *dst; + int num_encap = 0; + + *extended_dest = false; + if (!(fte->action.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST)) + return 0; + + list_for_each_entry(dst, &fte->node.children, node.list) { + if (dst->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_COUNTER) + continue; + if (dst->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_VPORT && + dst->dest_attr.vport.flags & MLX5_FLOW_DEST_VPORT_REFORMAT_ID) + num_encap++; + num_fwd_destinations++; + } + if (num_fwd_destinations > 1 && num_encap > 0) + *extended_dest = true; + + if (*extended_dest && !fw_log_max_fdb_encap_uplink) { + mlx5_core_warn(dev, "FW does not support extended destination"); + return -EOPNOTSUPP; + } + if (num_encap > (1 << fw_log_max_fdb_encap_uplink)) { + mlx5_core_warn(dev, "FW does not support more than %d encaps", + 1 << fw_log_max_fdb_encap_uplink); + return -EOPNOTSUPP; + } + + return 0; +} static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev, int opmod, int modify_mask, struct mlx5_flow_table *ft, unsigned group_id, struct fs_fte *fte) { - unsigned int inlen = MLX5_ST_SZ_BYTES(set_fte_in) + - fte->dests_size * MLX5_ST_SZ_BYTES(dest_format_struct); u32 out[MLX5_ST_SZ_DW(set_fte_out)] = {0}; + bool extended_dest = false; struct mlx5_flow_rule *dst; void *in_flow_context, *vlan; void *in_match_value; + unsigned int inlen; + int dst_cnt_size; void *in_dests; u32 *in; int err; + if (mlx5_set_extended_dest(dev, fte, &extended_dest)) + return -EOPNOTSUPP; + + if (!extended_dest) + dst_cnt_size = MLX5_ST_SZ_BYTES(dest_format_struct); + else + dst_cnt_size = MLX5_ST_SZ_BYTES(extended_dest_format); + + inlen = MLX5_ST_SZ_BYTES(set_fte_in) + fte->dests_size * dst_cnt_size; in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -343,9 +389,20 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev, MLX5_SET(flow_context, in_flow_context, group_id, group_id); MLX5_SET(flow_context, in_flow_context, flow_tag, fte->action.flow_tag); - MLX5_SET(flow_context, in_flow_context, action, fte->action.action); - MLX5_SET(flow_context, in_flow_context, packet_reformat_id, - fte->action.reformat_id); + MLX5_SET(flow_context, in_flow_context, extended_destination, + extended_dest); + if (extended_dest) { + u32 action; + + action = fte->action.action & + ~MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT; + MLX5_SET(flow_context, in_flow_context, action, action); + } else { + MLX5_SET(flow_context, in_flow_context, action, + fte->action.action); + MLX5_SET(flow_context, in_flow_context, packet_reformat_id, + fte->action.reformat_id); + } MLX5_SET(flow_context, in_flow_context, modify_header_id, fte->action.modify_id); @@ -392,6 +449,15 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev, MLX5_SET(dest_format_struct, in_dests, destination_eswitch_owner_vhca_id, dst->dest_attr.vport.vhca_id); + if (extended_dest) { + MLX5_SET(dest_format_struct, in_dests, + packet_reformat, + !!(dst->dest_attr.vport.flags & + MLX5_FLOW_DEST_VPORT_REFORMAT_ID)); + MLX5_SET(extended_dest_format, in_dests, + packet_reformat_id, + dst->dest_attr.vport.reformat_id); + } break; default: id = dst->dest_attr.tir_num; @@ -400,7 +466,7 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev, MLX5_SET(dest_format_struct, in_dests, destination_type, type); MLX5_SET(dest_format_struct, in_dests, destination_id, id); - in_dests += MLX5_ST_SZ_BYTES(dest_format_struct); + in_dests += dst_cnt_size; list_size++; } @@ -421,7 +487,7 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev, MLX5_SET(flow_counter_list, in_dests, flow_counter_id, dst->dest_attr.counter_id); - in_dests += MLX5_ST_SZ_BYTES(dest_format_struct); + in_dests += dst_cnt_size; list_size++; } if (list_size > max_list_size) { diff --git a/include/linux/mlx5/fs.h b/include/linux/mlx5/fs.h index 25ffd8018b72..9df51da04621 100644 --- a/include/linux/mlx5/fs.h +++ b/include/linux/mlx5/fs.h @@ -88,6 +88,7 @@ struct mlx5_flow_spec { enum { MLX5_FLOW_DEST_VPORT_VHCA_ID = BIT(0), + MLX5_FLOW_DEST_VPORT_REFORMAT_ID = BIT(1), }; struct mlx5_flow_destination { @@ -100,6 +101,7 @@ struct mlx5_flow_destination { struct { u16 num; u16 vhca_id; + u32 reformat_id; u8 flags; } vport; }; -- cgit v1.2.3-55-g7522 From 6c22a11957f46ca7e9b8db20ac7c6b05441c55ed Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Mon, 10 Dec 2018 13:15:17 -0800 Subject: net/mlx5: Remove the get protocol device interface entry This isn't used anywhere across the mlx5 driver stack, remove it. Signed-off-by: Or Gerlitz Reviewed-by: Saeed Mahameed Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/dev.c | 22 ---------------------- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 8 -------- include/linux/mlx5/driver.h | 2 -- 3 files changed, 32 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c index d2ed14bc37c3..ebc046fa97d3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c @@ -256,28 +256,6 @@ void mlx5_reload_interface(struct mlx5_core_dev *mdev, int protocol) mutex_unlock(&mlx5_intf_mutex); } -void *mlx5_get_protocol_dev(struct mlx5_core_dev *mdev, int protocol) -{ - struct mlx5_priv *priv = &mdev->priv; - struct mlx5_device_context *dev_ctx; - unsigned long flags; - void *result = NULL; - - spin_lock_irqsave(&priv->ctx_lock, flags); - - list_for_each_entry(dev_ctx, &mdev->priv.ctx_list, list) - if ((dev_ctx->intf->protocol == protocol) && - dev_ctx->intf->get_dev) { - result = dev_ctx->intf->get_dev(dev_ctx->context); - break; - } - - spin_unlock_irqrestore(&priv->ctx_lock, flags); - - return result; -} -EXPORT_SYMBOL(mlx5_get_protocol_dev); - /* Must be called with intf_mutex held */ void mlx5_add_dev_by_protocol(struct mlx5_core_dev *dev, int protocol) { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 56bc41b1c31f..a43092de3cc0 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -5160,20 +5160,12 @@ static void mlx5e_remove(struct mlx5_core_dev *mdev, void *vpriv) kfree(ppriv); } -static void *mlx5e_get_netdev(void *vpriv) -{ - struct mlx5e_priv *priv = vpriv; - - return priv->netdev; -} - static struct mlx5_interface mlx5e_interface = { .add = mlx5e_add, .remove = mlx5e_remove, .attach = mlx5e_attach, .detach = mlx5e_detach, .protocol = MLX5_INTERFACE_PROTOCOL_ETH, - .get_dev = mlx5e_get_netdev, }; void mlx5e_init(void) diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 584d8a5df7eb..cc29e880c733 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -1004,12 +1004,10 @@ struct mlx5_interface { void (*remove)(struct mlx5_core_dev *dev, void *context); int (*attach)(struct mlx5_core_dev *dev, void *context); void (*detach)(struct mlx5_core_dev *dev, void *context); - void * (*get_dev)(void *context); int protocol; struct list_head list; }; -void *mlx5_get_protocol_dev(struct mlx5_core_dev *mdev, int protocol); int mlx5_register_interface(struct mlx5_interface *intf); void mlx5_unregister_interface(struct mlx5_interface *intf); int mlx5_notifier_register(struct mlx5_core_dev *dev, struct notifier_block *nb); -- cgit v1.2.3-55-g7522 From ec1366c20721157cc2d077e84d2a062f93175ce5 Mon Sep 17 00:00:00 2001 From: Oz Shlomo Date: Thu, 25 Oct 2018 21:51:11 +0300 Subject: net/mlx5e: Store eswitch uplink representor state on a dedicated struct Currently only a single field in the representor private structure is relevant for uplink representors. As a pre-step to allow adding additional uplink representor fields, introduce uplink representor private structure. This is prepration step towards replacing egdev logic with the indirect block notification mechanism. This patch doesn't change any functionality. Signed-off-by: Oz Shlomo Reviewed-by: Eli Britstein Acked-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 6 +++--- drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | 9 ++++++++- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 2 +- 3 files changed, 12 insertions(+), 5 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index a137540b84bf..25a656b50264 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1244,7 +1244,7 @@ mlx5e_nic_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) { struct mlx5e_rep_priv *rpriv = mlx5e_rep_to_rep_priv(rep); struct mlx5e_priv *priv = netdev_priv(rpriv->netdev); - + struct mlx5_rep_uplink_priv *uplink_priv = &rpriv->uplink_priv; int err; if (test_bit(MLX5E_STATE_OPENED, &priv->state)) { @@ -1258,7 +1258,7 @@ mlx5e_nic_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) goto err_remove_sqs; /* init shared tc flow table */ - err = mlx5e_tc_esw_init(&rpriv->tc_ht); + err = mlx5e_tc_esw_init(&uplink_priv->tc_ht); if (err) goto err_neigh_cleanup; @@ -1281,7 +1281,7 @@ mlx5e_nic_rep_unload(struct mlx5_eswitch_rep *rep) mlx5e_remove_sqs_fwd_rules(priv); /* clean uplink offloaded TC rules, delete shared tc flow table */ - mlx5e_tc_esw_cleanup(&rpriv->tc_ht); + mlx5e_tc_esw_cleanup(&rpriv->uplink_priv.tc_ht); mlx5e_rep_neigh_cleanup(rpriv); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h index 844d32d5c29f..8d39c5c59aeb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h @@ -53,13 +53,20 @@ struct mlx5e_neigh_update_table { unsigned long min_interval; /* jiffies */ }; +struct mlx5_rep_uplink_priv { + /* Filters DB - instantiated by the uplink representor and shared by + * the uplink's VFs + */ + struct rhashtable tc_ht; +}; + struct mlx5e_rep_priv { struct mlx5_eswitch_rep *rep; struct mlx5e_neigh_update_table neigh_update; struct net_device *netdev; struct mlx5_flow_handle *vport_rx_rule; struct list_head vport_sqs_list; - struct rhashtable tc_ht; /* valid for uplink rep */ + struct mlx5_rep_uplink_priv uplink_priv; /* valid for uplink rep */ }; static inline diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index a3ff2492646d..42eb6ee37754 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -3018,7 +3018,7 @@ static struct rhashtable *get_tc_ht(struct mlx5e_priv *priv) if (MLX5_VPORT_MANAGER(priv->mdev) && esw->mode == SRIOV_OFFLOADS) { uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH); - return &uplink_rpriv->tc_ht; + return &uplink_rpriv->uplink_priv.tc_ht; } else return &priv->fs.tc.ht; } -- cgit v1.2.3-55-g7522 From f5bc2c5de1015caf7e49b0bf9e7e97077736045a Mon Sep 17 00:00:00 2001 From: Oz Shlomo Date: Sun, 28 Oct 2018 08:34:51 +0200 Subject: net/mlx5e: Support TC indirect block notifications for eswitch uplink reprs Towards using this mechanism as the means to offload tunnel decap rules set on SW tunnel devices instead of egdev, add the supporting structures and functions. Signed-off-by: Oz Shlomo Reviewed-by: Eli Britstein Reviewed-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 184 +++++++++++++++++++++++ drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | 13 ++ drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 10 ++ drivers/net/ethernet/mellanox/mlx5/core/en_tc.h | 3 + 4 files changed, 210 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index 25a656b50264..3aab3ce9119d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -49,6 +49,15 @@ static const char mlx5e_rep_driver_name[] = "mlx5e_rep"; +struct mlx5e_rep_indr_block_priv { + struct net_device *netdev; + struct mlx5e_rep_priv *rpriv; + + struct list_head list; +}; + +static void mlx5e_rep_indr_unregister_block(struct net_device *netdev); + static void mlx5e_rep_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *drvinfo) { @@ -518,6 +527,166 @@ static void mlx5e_rep_neigh_update(struct work_struct *work) neigh_release(n); } +static struct mlx5e_rep_indr_block_priv * +mlx5e_rep_indr_block_priv_lookup(struct mlx5e_rep_priv *rpriv, + struct net_device *netdev) +{ + struct mlx5e_rep_indr_block_priv *cb_priv; + + /* All callback list access should be protected by RTNL. */ + ASSERT_RTNL(); + + list_for_each_entry(cb_priv, + &rpriv->uplink_priv.tc_indr_block_priv_list, + list) + if (cb_priv->netdev == netdev) + return cb_priv; + + return NULL; +} + +static void mlx5e_rep_indr_clean_block_privs(struct mlx5e_rep_priv *rpriv) +{ + struct mlx5e_rep_indr_block_priv *cb_priv, *temp; + struct list_head *head = &rpriv->uplink_priv.tc_indr_block_priv_list; + + list_for_each_entry_safe(cb_priv, temp, head, list) { + mlx5e_rep_indr_unregister_block(cb_priv->netdev); + kfree(cb_priv); + } +} + +static int +mlx5e_rep_indr_offload(struct net_device *netdev, + struct tc_cls_flower_offload *flower, + struct mlx5e_rep_indr_block_priv *indr_priv) +{ + return -EOPNOTSUPP; +} + +static int mlx5e_rep_indr_setup_block_cb(enum tc_setup_type type, + void *type_data, void *indr_priv) +{ + struct mlx5e_rep_indr_block_priv *priv = indr_priv; + + switch (type) { + case TC_SETUP_CLSFLOWER: + return mlx5e_rep_indr_offload(priv->netdev, type_data, priv); + default: + return -EOPNOTSUPP; + } +} + +static int +mlx5e_rep_indr_setup_tc_block(struct net_device *netdev, + struct mlx5e_rep_priv *rpriv, + struct tc_block_offload *f) +{ + struct mlx5e_rep_indr_block_priv *indr_priv; + int err = 0; + + if (f->binder_type != TCF_BLOCK_BINDER_TYPE_CLSACT_INGRESS) + return -EOPNOTSUPP; + + switch (f->command) { + case TC_BLOCK_BIND: + indr_priv = mlx5e_rep_indr_block_priv_lookup(rpriv, netdev); + if (indr_priv) + return -EEXIST; + + indr_priv = kmalloc(sizeof(*indr_priv), GFP_KERNEL); + if (!indr_priv) + return -ENOMEM; + + indr_priv->netdev = netdev; + indr_priv->rpriv = rpriv; + list_add(&indr_priv->list, + &rpriv->uplink_priv.tc_indr_block_priv_list); + + err = tcf_block_cb_register(f->block, + mlx5e_rep_indr_setup_block_cb, + netdev, indr_priv, f->extack); + if (err) { + list_del(&indr_priv->list); + kfree(indr_priv); + } + + return err; + case TC_BLOCK_UNBIND: + tcf_block_cb_unregister(f->block, + mlx5e_rep_indr_setup_block_cb, + netdev); + indr_priv = mlx5e_rep_indr_block_priv_lookup(rpriv, netdev); + if (indr_priv) { + list_del(&indr_priv->list); + kfree(indr_priv); + } + + return 0; + default: + return -EOPNOTSUPP; + } + return 0; +} + +static +int mlx5e_rep_indr_setup_tc_cb(struct net_device *netdev, void *cb_priv, + enum tc_setup_type type, void *type_data) +{ + switch (type) { + case TC_SETUP_BLOCK: + return mlx5e_rep_indr_setup_tc_block(netdev, cb_priv, + type_data); + default: + return -EOPNOTSUPP; + } +} + +static int mlx5e_rep_indr_register_block(struct mlx5e_rep_priv *rpriv, + struct net_device *netdev) +{ + int err; + + err = __tc_indr_block_cb_register(netdev, rpriv, + mlx5e_rep_indr_setup_tc_cb, + netdev); + if (err) { + struct mlx5e_priv *priv = netdev_priv(rpriv->netdev); + + mlx5_core_err(priv->mdev, "Failed to register remote block notifier for %s err=%d\n", + netdev_name(netdev), err); + } + return err; +} + +static void mlx5e_rep_indr_unregister_block(struct net_device *netdev) +{ + __tc_indr_block_cb_unregister(netdev, mlx5e_rep_indr_setup_tc_cb, + netdev); +} + +static int mlx5e_nic_rep_netdevice_event(struct notifier_block *nb, + unsigned long event, void *ptr) +{ + struct mlx5e_rep_priv *rpriv = container_of(nb, struct mlx5e_rep_priv, + uplink_priv.netdevice_nb); + struct mlx5e_priv *priv = netdev_priv(rpriv->netdev); + struct net_device *netdev = netdev_notifier_info_to_dev(ptr); + + if (!mlx5e_tc_tun_device_to_offload(priv, netdev)) + return NOTIFY_OK; + + switch (event) { + case NETDEV_REGISTER: + mlx5e_rep_indr_register_block(rpriv, netdev); + break; + case NETDEV_UNREGISTER: + mlx5e_rep_indr_unregister_block(netdev); + break; + } + return NOTIFY_OK; +} + static struct mlx5e_neigh_hash_entry * mlx5e_rep_neigh_entry_lookup(struct mlx5e_priv *priv, struct mlx5e_neigh *m_neigh); @@ -1262,8 +1431,19 @@ mlx5e_nic_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) if (err) goto err_neigh_cleanup; + /* init indirect block notifications */ + INIT_LIST_HEAD(&uplink_priv->tc_indr_block_priv_list); + uplink_priv->netdevice_nb.notifier_call = mlx5e_nic_rep_netdevice_event; + err = register_netdevice_notifier(&uplink_priv->netdevice_nb); + if (err) { + mlx5_core_err(priv->mdev, "Failed to register netdev notifier\n"); + goto err_indirect_block_cleanup; + } + return 0; +err_indirect_block_cleanup: + mlx5e_tc_esw_cleanup(&uplink_priv->tc_ht); err_neigh_cleanup: mlx5e_rep_neigh_cleanup(rpriv); err_remove_sqs: @@ -1280,6 +1460,10 @@ mlx5e_nic_rep_unload(struct mlx5_eswitch_rep *rep) if (test_bit(MLX5E_STATE_OPENED, &priv->state)) mlx5e_remove_sqs_fwd_rules(priv); + /* clean indirect TC block notifications */ + unregister_netdevice_notifier(&rpriv->uplink_priv.netdevice_nb); + mlx5e_rep_indr_clean_block_privs(rpriv); + /* clean uplink offloaded TC rules, delete shared tc flow table */ mlx5e_tc_esw_cleanup(&rpriv->uplink_priv.tc_ht); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h index 8d39c5c59aeb..c078c6703dc7 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h @@ -58,6 +58,19 @@ struct mlx5_rep_uplink_priv { * the uplink's VFs */ struct rhashtable tc_ht; + + /* indirect block callbacks are invoked on bind/unbind events + * on registered higher level devices (e.g. tunnel devices) + * + * tc_indr_block_cb_priv_list is used to lookup indirect callback + * private data + * + * netdevice_nb is the netdev events notifier - used to register + * tunnel devices for block events + * + */ + struct list_head tc_indr_block_priv_list; + struct notifier_block netdevice_nb; }; struct mlx5e_rep_priv { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 42eb6ee37754..cb5bd3eaf738 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -2704,6 +2704,16 @@ out: return err; } +bool mlx5e_tc_tun_device_to_offload(struct mlx5e_priv *priv, + struct net_device *netdev) +{ + if (netif_is_vxlan(netdev) && + MLX5_CAP_ESW(priv->mdev, vxlan_encap_decap)) + return true; + + return false; +} + static int mlx5e_attach_encap(struct mlx5e_priv *priv, struct ip_tunnel_info *tun_info, struct net_device *mirred_dev, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h index 49436bf3b80a..378507988a32 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h @@ -70,6 +70,9 @@ void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe); int mlx5e_tc_num_filters(struct mlx5e_priv *priv); +bool mlx5e_tc_tun_device_to_offload(struct mlx5e_priv *priv, + struct net_device *netdev); + #else /* CONFIG_MLX5_ESWITCH */ static inline int mlx5e_tc_nic_init(struct mlx5e_priv *priv) { return 0; } static inline void mlx5e_tc_nic_cleanup(struct mlx5e_priv *priv) {} -- cgit v1.2.3-55-g7522 From 71d82d2a90327c9c816e65471b7859b601993f89 Mon Sep 17 00:00:00 2001 From: Oz Shlomo Date: Sun, 28 Oct 2018 09:14:50 +0200 Subject: net/mlx5e: Provide the TC filter netdev as parameter to flower callbacks Currently the driver controls flower filters that are installed on its devices. However, with the introduction of the indirect block notifications platform the driver may receive control events for filters that are installed on higher level net devices (e.g. tunnel devices). Therefore, the driver filter control API will not be able to implicitly assume the filter's net device. Explicitly specify the filter's net device, no functional change Signed-off-by: Oz Shlomo Reviewed-by: Eli Britstein Reviewed-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 9 ++++++--- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 9 ++++++--- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 6 +++--- drivers/net/ethernet/mellanox/mlx5/core/en_tc.h | 6 +++--- 4 files changed, 18 insertions(+), 12 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 06b1c0172a7b..88116a4750b0 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -3386,11 +3386,14 @@ static int mlx5e_setup_tc_cls_flower(struct mlx5e_priv *priv, { switch (cls_flower->command) { case TC_CLSFLOWER_REPLACE: - return mlx5e_configure_flower(priv, cls_flower, flags); + return mlx5e_configure_flower(priv->netdev, priv, cls_flower, + flags); case TC_CLSFLOWER_DESTROY: - return mlx5e_delete_flower(priv, cls_flower, flags); + return mlx5e_delete_flower(priv->netdev, priv, cls_flower, + flags); case TC_CLSFLOWER_STATS: - return mlx5e_stats_flower(priv, cls_flower, flags); + return mlx5e_stats_flower(priv->netdev, priv, cls_flower, + flags); default: return -EOPNOTSUPP; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index 3aab3ce9119d..ab9c7ec77520 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1007,11 +1007,14 @@ mlx5e_rep_setup_tc_cls_flower(struct mlx5e_priv *priv, { switch (cls_flower->command) { case TC_CLSFLOWER_REPLACE: - return mlx5e_configure_flower(priv, cls_flower, flags); + return mlx5e_configure_flower(priv->netdev, priv, cls_flower, + flags); case TC_CLSFLOWER_DESTROY: - return mlx5e_delete_flower(priv, cls_flower, flags); + return mlx5e_delete_flower(priv->netdev, priv, cls_flower, + flags); case TC_CLSFLOWER_STATS: - return mlx5e_stats_flower(priv, cls_flower, flags); + return mlx5e_stats_flower(priv->netdev, priv, cls_flower, + flags); default: return -EOPNOTSUPP; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index cb5bd3eaf738..b614513df127 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -3178,7 +3178,7 @@ mlx5e_tc_add_flow(struct mlx5e_priv *priv, return err; } -int mlx5e_configure_flower(struct mlx5e_priv *priv, +int mlx5e_configure_flower(struct net_device *dev, struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, int flags) { struct netlink_ext_ack *extack = f->common.extack; @@ -3224,7 +3224,7 @@ static bool same_flow_direction(struct mlx5e_tc_flow *flow, int flags) return false; } -int mlx5e_delete_flower(struct mlx5e_priv *priv, +int mlx5e_delete_flower(struct net_device *dev, struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, int flags) { struct rhashtable *tc_ht = get_tc_ht(priv); @@ -3243,7 +3243,7 @@ int mlx5e_delete_flower(struct mlx5e_priv *priv, return 0; } -int mlx5e_stats_flower(struct mlx5e_priv *priv, +int mlx5e_stats_flower(struct net_device *dev, struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, int flags) { struct rhashtable *tc_ht = get_tc_ht(priv); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h index 378507988a32..004e679a4f53 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h @@ -51,12 +51,12 @@ void mlx5e_tc_nic_cleanup(struct mlx5e_priv *priv); int mlx5e_tc_esw_init(struct rhashtable *tc_ht); void mlx5e_tc_esw_cleanup(struct rhashtable *tc_ht); -int mlx5e_configure_flower(struct mlx5e_priv *priv, +int mlx5e_configure_flower(struct net_device *dev, struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, int flags); -int mlx5e_delete_flower(struct mlx5e_priv *priv, +int mlx5e_delete_flower(struct net_device *dev, struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, int flags); -int mlx5e_stats_flower(struct mlx5e_priv *priv, +int mlx5e_stats_flower(struct net_device *dev, struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, int flags); struct mlx5e_encap_entry; -- cgit v1.2.3-55-g7522 From d11afc2600377c6775abcbc1fd410f199c5bf458 Mon Sep 17 00:00:00 2001 From: Oz Shlomo Date: Sun, 28 Oct 2018 10:46:34 +0200 Subject: net/mlx5e: Propagate the filter's net device to mlx5e structures Propagate the filter's net_device parameter to the tc flower parsed attributes structure so that it can later be used in tunnel decap offloading sequences. Pre-step for replacing egdev logic with the indirect block notification mechanism. Signed-off-by: Oz Shlomo Reviewed-by: Eli Britstein Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 26 ++++++++++++++++++------- 1 file changed, 19 insertions(+), 7 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index b614513df127..3271eda8098b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -96,6 +96,7 @@ struct mlx5e_tc_flow { struct mlx5e_tc_flow_parse_attr { struct ip_tunnel_info tun_info; + struct net_device *filter_dev; struct mlx5_flow_spec spec; int num_mod_hdr_actions; void *mod_hdr_actions; @@ -3054,10 +3055,6 @@ mlx5e_alloc_flow(struct mlx5e_priv *priv, int attr_size, flow->flags = flow_flags; flow->priv = priv; - err = parse_cls_flower(priv, flow, &parse_attr->spec, f); - if (err) - goto err_free; - *__flow = flow; *__parse_attr = parse_attr; @@ -3073,6 +3070,7 @@ static int mlx5e_add_fdb_flow(struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, u16 flow_flags, + struct net_device *filter_dev, struct mlx5e_tc_flow **__flow) { struct netlink_ext_ack *extack = f->common.extack; @@ -3086,6 +3084,11 @@ mlx5e_add_fdb_flow(struct mlx5e_priv *priv, &parse_attr, &flow); if (err) goto out; + parse_attr->filter_dev = filter_dev; + flow->esw_attr->parse_attr = parse_attr; + err = parse_cls_flower(flow->priv, flow, &parse_attr->spec, f); + if (err) + goto err_free; flow->esw_attr->chain = f->common.chain_index; flow->esw_attr->prio = TC_H_MAJ(f->common.prio) >> 16; @@ -3116,6 +3119,7 @@ static int mlx5e_add_nic_flow(struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, u16 flow_flags, + struct net_device *filter_dev, struct mlx5e_tc_flow **__flow) { struct netlink_ext_ack *extack = f->common.extack; @@ -3134,6 +3138,11 @@ mlx5e_add_nic_flow(struct mlx5e_priv *priv, if (err) goto out; + parse_attr->filter_dev = filter_dev; + err = parse_cls_flower(flow->priv, flow, &parse_attr->spec, f); + if (err) + goto err_free; + err = parse_tc_nic_actions(priv, f->exts, parse_attr, flow, extack); if (err) goto err_free; @@ -3159,6 +3168,7 @@ static int mlx5e_tc_add_flow(struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, int flags, + struct net_device *filter_dev, struct mlx5e_tc_flow **flow) { struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; @@ -3171,9 +3181,11 @@ mlx5e_tc_add_flow(struct mlx5e_priv *priv, return -EOPNOTSUPP; if (esw && esw->mode == SRIOV_OFFLOADS) - err = mlx5e_add_fdb_flow(priv, f, flow_flags, flow); + err = mlx5e_add_fdb_flow(priv, f, flow_flags, + filter_dev, flow); else - err = mlx5e_add_nic_flow(priv, f, flow_flags, flow); + err = mlx5e_add_nic_flow(priv, f, flow_flags, + filter_dev, flow); return err; } @@ -3196,7 +3208,7 @@ int mlx5e_configure_flower(struct net_device *dev, struct mlx5e_priv *priv, goto out; } - err = mlx5e_tc_add_flow(priv, f, flags, &flow); + err = mlx5e_tc_add_flow(priv, f, flags, dev, &flow); if (err) goto out; -- cgit v1.2.3-55-g7522 From ef381359e3a81422cee85ddee1f7dc3ac90f3c38 Mon Sep 17 00:00:00 2001 From: Oz Shlomo Date: Sun, 28 Oct 2018 13:03:54 +0200 Subject: net/mlx5e: Replace egdev with indirect block notifications Use TC indirect block notifications to offload filters that are configured on higher level device interfaces (e.g. tunnel devices). This mechanism replaces the current egdev implementation. Signed-off-by: Oz Shlomo Reviewed-by: Eli Britstein Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 47 +++++++++++------------- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 14 +++++++ 2 files changed, 36 insertions(+), 25 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index ab9c7ec77520..5cac4de435c9 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -561,7 +561,27 @@ mlx5e_rep_indr_offload(struct net_device *netdev, struct tc_cls_flower_offload *flower, struct mlx5e_rep_indr_block_priv *indr_priv) { - return -EOPNOTSUPP; + int err = 0; + struct mlx5e_priv *priv = netdev_priv(indr_priv->rpriv->netdev); + + switch (flower->command) { + case TC_CLSFLOWER_REPLACE: + err = mlx5e_configure_flower(netdev, priv, + flower, MLX5E_TC_EGRESS); + break; + case TC_CLSFLOWER_DESTROY: + err = mlx5e_delete_flower(netdev, priv, + flower, MLX5E_TC_EGRESS); + break; + case TC_CLSFLOWER_STATS: + err = mlx5e_stats_flower(netdev, priv, + flower, MLX5E_TC_EGRESS); + break; + default: + err = -EOPNOTSUPP; + } + + return err; } static int mlx5e_rep_indr_setup_block_cb(enum tc_setup_type type, @@ -1020,19 +1040,6 @@ mlx5e_rep_setup_tc_cls_flower(struct mlx5e_priv *priv, } } -static int mlx5e_rep_setup_tc_cb_egdev(enum tc_setup_type type, void *type_data, - void *cb_priv) -{ - struct mlx5e_priv *priv = cb_priv; - - switch (type) { - case TC_SETUP_CLSFLOWER: - return mlx5e_rep_setup_tc_cls_flower(priv, type_data, MLX5E_TC_EGRESS); - default: - return -EOPNOTSUPP; - } -} - static int mlx5e_rep_setup_tc_cb(enum tc_setup_type type, void *type_data, void *cb_priv) { @@ -1516,24 +1523,16 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) uplink_rpriv = mlx5_eswitch_get_uplink_priv(dev->priv.eswitch, REP_ETH); upriv = netdev_priv(uplink_rpriv->netdev); - err = tc_setup_cb_egdev_register(netdev, mlx5e_rep_setup_tc_cb_egdev, - upriv); - if (err) - goto err_neigh_cleanup; err = register_netdev(netdev); if (err) { pr_warn("Failed to register representor netdev for vport %d\n", rep->vport); - goto err_egdev_cleanup; + goto err_neigh_cleanup; } return 0; -err_egdev_cleanup: - tc_setup_cb_egdev_unregister(netdev, mlx5e_rep_setup_tc_cb_egdev, - upriv); - err_neigh_cleanup: mlx5e_rep_neigh_cleanup(rpriv); @@ -1560,8 +1559,6 @@ mlx5e_vport_rep_unload(struct mlx5_eswitch_rep *rep) uplink_rpriv = mlx5_eswitch_get_uplink_priv(priv->mdev->priv.eswitch, REP_ETH); upriv = netdev_priv(uplink_rpriv->netdev); - tc_setup_cb_egdev_unregister(netdev, mlx5e_rep_setup_tc_cb_egdev, - upriv); mlx5e_rep_neigh_cleanup(rpriv); mlx5e_detach_netdev(priv); mlx5e_destroy_netdev(priv); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 3271eda8098b..6ca94a2f167f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -2906,6 +2906,13 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, struct net_device *out_dev; out_dev = tcf_mirred_dev(a); + if (!out_dev) { + /* out_dev is NULL when filters with + * non-existing mirred device are replayed to + * the driver. + */ + return -EINVAL; + } if (attr->out_count >= MLX5_MAX_FLOW_FWD_VPORTS) { NL_SET_ERR_MSG_MOD(extack, @@ -2932,6 +2939,13 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, MLX5_FLOW_CONTEXT_ACTION_FWD_DEST | MLX5_FLOW_CONTEXT_ACTION_COUNT; /* attr->out_rep is resolved when we handle encap */ + } else if (parse_attr->filter_dev != priv->netdev) { + /* All mlx5 devices are called to configure + * high level device filters. Therefore, the + * *attempt* to install a filter on invalid + * eswitch should not trigger an explicit error + */ + return -EINVAL; } else { NL_SET_ERR_MSG_MOD(extack, "devices are not on same switch HW, can't offload forwarding"); -- cgit v1.2.3-55-g7522 From ea7162ac3a33a44a7bb0b5793057015d25617e68 Mon Sep 17 00:00:00 2001 From: Oz Shlomo Date: Wed, 14 Nov 2018 15:41:50 +0200 Subject: net/mlx5e: Refactor VXLAN tunnel encap offloading code Separates the vxlan header encap logic from the general ipv4/6 encapsulation methods, thus allowing the common IP encap/decap code to branch in downstream patch to multiple IP tunnels. Code refactoring with no functional change. Signed-off-by: Oz Shlomo Reviewed-by: Eli Britstein Reviewed-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 189 +++++++++--------------- 1 file changed, 70 insertions(+), 119 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 6ca94a2f167f..ed04ab8287e9 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -2416,70 +2416,25 @@ static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv, return 0; } -static void gen_vxlan_header_ipv4(struct net_device *out_dev, - char buf[], int encap_size, - unsigned char h_dest[ETH_ALEN], - u8 tos, u8 ttl, - __be32 daddr, - __be32 saddr, - __be16 udp_dst_port, - __be32 vx_vni) -{ - struct ethhdr *eth = (struct ethhdr *)buf; - struct iphdr *ip = (struct iphdr *)((char *)eth + sizeof(struct ethhdr)); - struct udphdr *udp = (struct udphdr *)((char *)ip + sizeof(struct iphdr)); - struct vxlanhdr *vxh = (struct vxlanhdr *)((char *)udp + sizeof(struct udphdr)); - - memset(buf, 0, encap_size); - - ether_addr_copy(eth->h_dest, h_dest); - ether_addr_copy(eth->h_source, out_dev->dev_addr); - eth->h_proto = htons(ETH_P_IP); - - ip->daddr = daddr; - ip->saddr = saddr; - - ip->tos = tos; - ip->ttl = ttl; - ip->protocol = IPPROTO_UDP; - ip->version = 0x4; - ip->ihl = 0x5; +static int mlx5e_gen_vxlan_header(char buf[], struct ip_tunnel_key *tun_key) +{ + __be32 tun_id = tunnel_id_to_key32(tun_key->tun_id); + struct udphdr *udp = (struct udphdr *)(buf); + struct vxlanhdr *vxh = (struct vxlanhdr *) + ((char *)udp + sizeof(struct udphdr)); - udp->dest = udp_dst_port; + udp->dest = tun_key->tp_dst; vxh->vx_flags = VXLAN_HF_VNI; - vxh->vx_vni = vxlan_vni_field(vx_vni); + vxh->vx_vni = vxlan_vni_field(tun_id); + + return 0; } -static void gen_vxlan_header_ipv6(struct net_device *out_dev, - char buf[], int encap_size, - unsigned char h_dest[ETH_ALEN], - u8 tos, u8 ttl, - struct in6_addr *daddr, - struct in6_addr *saddr, - __be16 udp_dst_port, - __be32 vx_vni) +static int mlx5e_gen_ip_tunnel_header(char buf[], __u8 *ip_proto, + struct ip_tunnel_key *tun_key) { - struct ethhdr *eth = (struct ethhdr *)buf; - struct ipv6hdr *ip6h = (struct ipv6hdr *)((char *)eth + sizeof(struct ethhdr)); - struct udphdr *udp = (struct udphdr *)((char *)ip6h + sizeof(struct ipv6hdr)); - struct vxlanhdr *vxh = (struct vxlanhdr *)((char *)udp + sizeof(struct udphdr)); - - memset(buf, 0, encap_size); - - ether_addr_copy(eth->h_dest, h_dest); - ether_addr_copy(eth->h_source, out_dev->dev_addr); - eth->h_proto = htons(ETH_P_IPV6); - - ip6_flow_hdr(ip6h, tos, 0); - /* the HW fills up ipv6 payload len */ - ip6h->nexthdr = IPPROTO_UDP; - ip6h->hop_limit = ttl; - ip6h->daddr = *daddr; - ip6h->saddr = *saddr; - - udp->dest = udp_dst_port; - vxh->vx_flags = VXLAN_HF_VNI; - vxh->vx_vni = vxlan_vni_field(vx_vni); + *ip_proto = IPPROTO_UDP; + return mlx5e_gen_vxlan_header(buf, tun_key); } static int mlx5e_create_encap_header_ipv4(struct mlx5e_priv *priv, @@ -2487,13 +2442,17 @@ static int mlx5e_create_encap_header_ipv4(struct mlx5e_priv *priv, struct mlx5e_encap_entry *e) { int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size); - int ipv4_encap_size = ETH_HLEN + sizeof(struct iphdr) + VXLAN_HLEN; + int ipv4_encap_size = ETH_HLEN + + sizeof(struct iphdr) + + VXLAN_HLEN; struct ip_tunnel_key *tun_key = &e->tun_info.key; struct net_device *out_dev; struct neighbour *n = NULL; struct flowi4 fl4 = {}; - u8 nud_state, tos, ttl; char *encap_header; + struct ethhdr *eth; + u8 nud_state, ttl; + struct iphdr *ip; int err; if (max_encap_size < ipv4_encap_size) { @@ -2506,22 +2465,11 @@ static int mlx5e_create_encap_header_ipv4(struct mlx5e_priv *priv, if (!encap_header) return -ENOMEM; - switch (e->tunnel_type) { - case MLX5_REFORMAT_TYPE_L2_TO_VXLAN: - fl4.flowi4_proto = IPPROTO_UDP; - fl4.fl4_dport = tun_key->tp_dst; - break; - default: - err = -EOPNOTSUPP; - goto free_encap; - } - - tos = tun_key->tos; - ttl = tun_key->ttl; - + /* add the IP fields */ fl4.flowi4_tos = tun_key->tos; fl4.daddr = tun_key->u.ipv4.dst; fl4.saddr = tun_key->u.ipv4.src; + ttl = tun_key->ttl; err = mlx5e_route_lookup_ipv4(priv, mirred_dev, &out_dev, &fl4, &n, &ttl); @@ -2536,7 +2484,7 @@ static int mlx5e_create_encap_header_ipv4(struct mlx5e_priv *priv, memcpy(&e->m_neigh.dst_ip, n->primary_key, n->tbl->key_len); e->out_dev = out_dev; - /* It's importent to add the neigh to the hash table before checking + /* It's important to add the neigh to the hash table before checking * the neigh validity state. So if we'll get a notification, in case the * neigh changes it's validity state, we would find the relevant neigh * in the hash. @@ -2550,18 +2498,27 @@ static int mlx5e_create_encap_header_ipv4(struct mlx5e_priv *priv, ether_addr_copy(e->h_dest, n->ha); read_unlock_bh(&n->lock); - switch (e->tunnel_type) { - case MLX5_REFORMAT_TYPE_L2_TO_VXLAN: - gen_vxlan_header_ipv4(out_dev, encap_header, - ipv4_encap_size, e->h_dest, tos, ttl, - fl4.daddr, - fl4.saddr, tun_key->tp_dst, - tunnel_id_to_key32(tun_key->tun_id)); - break; - default: - err = -EOPNOTSUPP; + /* add ethernet header */ + eth = (struct ethhdr *)encap_header; + ether_addr_copy(eth->h_dest, e->h_dest); + ether_addr_copy(eth->h_source, out_dev->dev_addr); + eth->h_proto = htons(ETH_P_IP); + + /* add ip header */ + ip = (struct iphdr *)((char *)eth + sizeof(struct ethhdr)); + ip->tos = tun_key->tos; + ip->version = 0x4; + ip->ihl = 0x5; + ip->ttl = ttl; + ip->daddr = fl4.daddr; + ip->saddr = fl4.saddr; + + /* add tunneling protocol header */ + err = mlx5e_gen_ip_tunnel_header((char *)ip + sizeof(struct iphdr), + &ip->protocol, tun_key); + if (err) goto destroy_neigh_entry; - } + e->encap_size = ipv4_encap_size; e->encap_header = encap_header; @@ -2598,13 +2555,17 @@ static int mlx5e_create_encap_header_ipv6(struct mlx5e_priv *priv, struct mlx5e_encap_entry *e) { int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size); - int ipv6_encap_size = ETH_HLEN + sizeof(struct ipv6hdr) + VXLAN_HLEN; + int ipv6_encap_size = ETH_HLEN + + sizeof(struct ipv6hdr) + + VXLAN_HLEN; struct ip_tunnel_key *tun_key = &e->tun_info.key; struct net_device *out_dev; struct neighbour *n = NULL; struct flowi6 fl6 = {}; - u8 nud_state, tos, ttl; + struct ipv6hdr *ip6h; char *encap_header; + struct ethhdr *eth; + u8 nud_state, ttl; int err; if (max_encap_size < ipv6_encap_size) { @@ -2617,17 +2578,6 @@ static int mlx5e_create_encap_header_ipv6(struct mlx5e_priv *priv, if (!encap_header) return -ENOMEM; - switch (e->tunnel_type) { - case MLX5_REFORMAT_TYPE_L2_TO_VXLAN: - fl6.flowi6_proto = IPPROTO_UDP; - fl6.fl6_dport = tun_key->tp_dst; - break; - default: - err = -EOPNOTSUPP; - goto free_encap; - } - - tos = tun_key->tos; ttl = tun_key->ttl; fl6.flowlabel = ip6_make_flowinfo(RT_TOS(tun_key->tos), tun_key->label); @@ -2661,18 +2611,25 @@ static int mlx5e_create_encap_header_ipv6(struct mlx5e_priv *priv, ether_addr_copy(e->h_dest, n->ha); read_unlock_bh(&n->lock); - switch (e->tunnel_type) { - case MLX5_REFORMAT_TYPE_L2_TO_VXLAN: - gen_vxlan_header_ipv6(out_dev, encap_header, - ipv6_encap_size, e->h_dest, tos, ttl, - &fl6.daddr, - &fl6.saddr, tun_key->tp_dst, - tunnel_id_to_key32(tun_key->tun_id)); - break; - default: - err = -EOPNOTSUPP; + /* add ethernet header */ + eth = (struct ethhdr *)encap_header; + ether_addr_copy(eth->h_dest, e->h_dest); + ether_addr_copy(eth->h_source, out_dev->dev_addr); + eth->h_proto = htons(ETH_P_IPV6); + + /* add ip header */ + ip6h = (struct ipv6hdr *)((char *)eth + sizeof(struct ethhdr)); + ip6_flow_hdr(ip6h, tun_key->tos, 0); + /* the HW fills up ipv6 payload len */ + ip6h->hop_limit = ttl; + ip6h->daddr = fl6.daddr; + ip6h->saddr = fl6.saddr; + + /* add tunneling protocol header */ + err = mlx5e_gen_ip_tunnel_header((char *)ip6h + sizeof(struct ipv6hdr), + &ip6h->nexthdr, tun_key); + if (err) goto destroy_neigh_entry; - } e->encap_size = ipv6_encap_size; e->encap_header = encap_header; @@ -2731,17 +2688,11 @@ static int mlx5e_attach_encap(struct mlx5e_priv *priv, uintptr_t hash_key; bool found = false; - /* udp dst port must be set */ - if (!memchr_inv(&key->tp_dst, 0, sizeof(key->tp_dst))) - goto vxlan_encap_offload_err; - - /* setting udp src port isn't supported */ - if (memchr_inv(&key->tp_src, 0, sizeof(key->tp_src))) { -vxlan_encap_offload_err: + if (!MLX5_CAP_ESW(priv->mdev, vxlan_encap_decap)) { NL_SET_ERR_MSG_MOD(extack, - "must set udp dst port and not set udp src port"); + "vxlan HW offloading is not supported"); netdev_warn(priv->netdev, - "must set udp dst port and not set udp src port\n"); + "vxlan HW offloading is not supported\n"); return -EOPNOTSUPP; } -- cgit v1.2.3-55-g7522 From 4d70564d1c9e812b83f30c8dc23cd506abfcfc08 Mon Sep 17 00:00:00 2001 From: Oz Shlomo Date: Wed, 14 Nov 2018 15:41:50 +0200 Subject: net/mlx5e: Refactor VXLAN tunnel decap offloading code Separates the vxlan header match handling from the matching on the general fields of ipv4/6 tunnels, thus allowing the common IP tunnel match code to branch in down stream patch, to multiple IP tunnels. This patch doesn't add any functionality. Signed-off-by: Oz Shlomo Reviewed-by: Eli Britstein Reviewed-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 90 ++++++++++++++----------- 1 file changed, 51 insertions(+), 39 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index ed04ab8287e9..74823cab1d16 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -1193,21 +1193,59 @@ static void mlx5e_tc_del_flow(struct mlx5e_priv *priv, mlx5e_tc_del_nic_flow(priv, flow); } -static void parse_vxlan_attr(struct mlx5_flow_spec *spec, - struct tc_cls_flower_offload *f) +static int parse_tunnel_vxlan_attr(struct mlx5e_priv *priv, + struct mlx5_flow_spec *spec, + struct tc_cls_flower_offload *f, + void *headers_c, + void *headers_v) { - void *headers_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, - outer_headers); - void *headers_v = MLX5_ADDR_OF(fte_match_param, spec->match_value, - outer_headers); - void *misc_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, + struct netlink_ext_ack *extack = f->common.extack; + struct flow_dissector_key_ports *key = + skb_flow_dissector_target(f->dissector, + FLOW_DISSECTOR_KEY_ENC_PORTS, + f->key); + struct flow_dissector_key_ports *mask = + skb_flow_dissector_target(f->dissector, + FLOW_DISSECTOR_KEY_ENC_PORTS, + f->mask); + void *misc_c = MLX5_ADDR_OF(fte_match_param, + spec->match_criteria, misc_parameters); - void *misc_v = MLX5_ADDR_OF(fte_match_param, spec->match_value, + void *misc_v = MLX5_ADDR_OF(fte_match_param, + spec->match_value, misc_parameters); + /* Full udp dst port must be given */ + if (!dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_PORTS) || + memchr_inv(&mask->dst, 0xff, sizeof(mask->dst))) { + NL_SET_ERR_MSG_MOD(extack, + "VXLAN decap filter must include enc_dst_port condition"); + netdev_warn(priv->netdev, + "VXLAN decap filter must include enc_dst_port condition\n"); + return -EOPNOTSUPP; + } + + /* udp dst port must be knonwn as a VXLAN port */ + if (!mlx5_vxlan_lookup_port(priv->mdev->vxlan, be16_to_cpu(key->dst))) { + NL_SET_ERR_MSG_MOD(extack, + "Matched UDP port is not registered as a VXLAN port"); + netdev_warn(priv->netdev, + "UDP port %d is not registered as a VXLAN port\n", + be16_to_cpu(key->dst)); + return -EOPNOTSUPP; + } + + /* dst UDP port is valid here */ MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, headers_c, ip_protocol); MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_protocol, IPPROTO_UDP); + MLX5_SET(fte_match_set_lyr_2_4, headers_c, udp_dport, ntohs(mask->dst)); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, udp_dport, ntohs(key->dst)); + + MLX5_SET(fte_match_set_lyr_2_4, headers_c, udp_sport, ntohs(mask->src)); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, udp_sport, ntohs(key->src)); + + /* match on VNI */ if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_KEYID)) { struct flow_dissector_key_keyid *key = skb_flow_dissector_target(f->dissector, @@ -1222,6 +1260,7 @@ static void parse_vxlan_attr(struct mlx5_flow_spec *spec, MLX5_SET(fte_match_set_misc, misc_v, vxlan_vni, be32_to_cpu(key->keyid)); } + return 0; } static int parse_tunnel_attr(struct mlx5e_priv *priv, @@ -1240,39 +1279,12 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv, f->key); if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_PORTS)) { - struct flow_dissector_key_ports *key = - skb_flow_dissector_target(f->dissector, - FLOW_DISSECTOR_KEY_ENC_PORTS, - f->key); - struct flow_dissector_key_ports *mask = - skb_flow_dissector_target(f->dissector, - FLOW_DISSECTOR_KEY_ENC_PORTS, - f->mask); + int err = 0; - /* Full udp dst port must be given */ - if (memchr_inv(&mask->dst, 0xff, sizeof(mask->dst))) + err = parse_tunnel_vxlan_attr(priv, spec, f, + headers_c, headers_v); + if (err) goto vxlan_match_offload_err; - - if (mlx5_vxlan_lookup_port(priv->mdev->vxlan, be16_to_cpu(key->dst)) && - MLX5_CAP_ESW(priv->mdev, vxlan_encap_decap)) - parse_vxlan_attr(spec, f); - else { - NL_SET_ERR_MSG_MOD(extack, - "port isn't an offloaded vxlan udp dport"); - netdev_warn(priv->netdev, - "%d isn't an offloaded vxlan udp dport\n", be16_to_cpu(key->dst)); - return -EOPNOTSUPP; - } - - MLX5_SET(fte_match_set_lyr_2_4, headers_c, - udp_dport, ntohs(mask->dst)); - MLX5_SET(fte_match_set_lyr_2_4, headers_v, - udp_dport, ntohs(key->dst)); - - MLX5_SET(fte_match_set_lyr_2_4, headers_c, - udp_sport, ntohs(mask->src)); - MLX5_SET(fte_match_set_lyr_2_4, headers_v, - udp_sport, ntohs(key->src)); } else { /* udp dst port must be given */ vxlan_match_offload_err: NL_SET_ERR_MSG_MOD(extack, -- cgit v1.2.3-55-g7522 From 54c177ca9c6efe5df516eefb886761b89a82eaf0 Mon Sep 17 00:00:00 2001 From: Oz Shlomo Date: Wed, 14 Nov 2018 15:21:27 +0200 Subject: net/mlx5e: Branch according to classified tunnel type Currently the tunnel offloading encap/decap methods assumes that VXLAN is the sole tunneling protocol. Lay the infrastructure for supporting multiple tunneling protocols by branching according to the tunnel net device kind. Encap filters tunnel type is determined according to the egress/mirred net device. Decap filters classify the tunnel type according to the filter's ingress net device kind. Distinguish between the tunnel type as defined by the SW model and the FW reformat type that specifies the HW operation being made. Signed-off-by: Oz Shlomo Reviewed-by: Eli Britstein Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | 2 + drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 160 ++++++++++++++++------- 2 files changed, 114 insertions(+), 48 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h index c078c6703dc7..2c8798332c32 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h @@ -149,6 +149,8 @@ struct mlx5e_encap_entry { struct net_device *out_dev; int tunnel_type; + int tunnel_hlen; + int reformat_type; u8 flags; char *encap_header; int encap_size; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 74823cab1d16..7d7f490d8ff1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -79,6 +79,13 @@ enum { #define MLX5E_TC_MAX_SPLITS 1 +enum { + MLX5E_TC_TUNNEL_TYPE_UNKNOWN, + MLX5E_TC_TUNNEL_TYPE_VXLAN +}; + +static int mlx5e_get_tunnel_type(struct net_device *tunnel_dev); + struct mlx5e_tc_flow { struct rhash_head node; struct mlx5e_priv *priv; @@ -677,6 +684,14 @@ static void mlx5e_hairpin_flow_del(struct mlx5e_priv *priv, } } +static const char *mlx5e_netdev_kind(struct net_device *dev) +{ + if (dev->rtnl_link_ops) + return dev->rtnl_link_ops->kind; + else + return ""; +} + static int mlx5e_tc_add_nic_flow(struct mlx5e_priv *priv, struct mlx5e_tc_flow_parse_attr *parse_attr, @@ -1037,7 +1052,8 @@ void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv, struct mlx5e_tc_flow *flow; int err; - err = mlx5_packet_reformat_alloc(priv->mdev, e->tunnel_type, + err = mlx5_packet_reformat_alloc(priv->mdev, + e->reformat_type, e->encap_size, e->encap_header, MLX5_FLOW_NAMESPACE_FDB, &e->encap_id); @@ -1265,7 +1281,8 @@ static int parse_tunnel_vxlan_attr(struct mlx5e_priv *priv, static int parse_tunnel_attr(struct mlx5e_priv *priv, struct mlx5_flow_spec *spec, - struct tc_cls_flower_offload *f) + struct tc_cls_flower_offload *f, + struct net_device *filter_dev) { struct netlink_ext_ack *extack = f->common.extack; void *headers_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, @@ -1277,20 +1294,28 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv, skb_flow_dissector_target(f->dissector, FLOW_DISSECTOR_KEY_ENC_CONTROL, f->key); + int tunnel_type; + int err = 0; - if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_PORTS)) { - int err = 0; - + tunnel_type = mlx5e_get_tunnel_type(filter_dev); + if (tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN) { err = parse_tunnel_vxlan_attr(priv, spec, f, headers_c, headers_v); - if (err) - goto vxlan_match_offload_err; - } else { /* udp dst port must be given */ -vxlan_match_offload_err: + } else { NL_SET_ERR_MSG_MOD(extack, - "IP tunnel decap offload supported only for vxlan, must set UDP dport"); + "decapsulation offload is not supported"); netdev_warn(priv->netdev, - "IP tunnel decap offload supported only for vxlan, must set UDP dport\n"); + "decapsulation offload is not supported for %s net device (%d)\n", + mlx5e_netdev_kind(filter_dev), tunnel_type); + return -EOPNOTSUPP; + } + + if (err) { + NL_SET_ERR_MSG_MOD(extack, + "failed to parse tunnel attributes"); + netdev_warn(priv->netdev, + "failed to parse %s tunnel attributes (%d)\n", + mlx5e_netdev_kind(filter_dev), tunnel_type); return -EOPNOTSUPP; } @@ -1395,6 +1420,7 @@ vxlan_match_offload_err: static int __parse_cls_flower(struct mlx5e_priv *priv, struct mlx5_flow_spec *spec, struct tc_cls_flower_offload *f, + struct net_device *filter_dev, u8 *match_level) { struct netlink_ext_ack *extack = f->common.extack; @@ -1446,7 +1472,7 @@ static int __parse_cls_flower(struct mlx5e_priv *priv, switch (key->addr_type) { case FLOW_DISSECTOR_KEY_IPV4_ADDRS: case FLOW_DISSECTOR_KEY_IPV6_ADDRS: - if (parse_tunnel_attr(priv, spec, f)) + if (parse_tunnel_attr(priv, spec, f, filter_dev)) return -EOPNOTSUPP; break; default: @@ -1788,7 +1814,8 @@ static int __parse_cls_flower(struct mlx5e_priv *priv, static int parse_cls_flower(struct mlx5e_priv *priv, struct mlx5e_tc_flow *flow, struct mlx5_flow_spec *spec, - struct tc_cls_flower_offload *f) + struct tc_cls_flower_offload *f, + struct net_device *filter_dev) { struct netlink_ext_ack *extack = f->common.extack; struct mlx5_core_dev *dev = priv->mdev; @@ -1798,7 +1825,7 @@ static int parse_cls_flower(struct mlx5e_priv *priv, u8 match_level; int err; - err = __parse_cls_flower(priv, spec, f, &match_level); + err = __parse_cls_flower(priv, spec, f, filter_dev, &match_level); if (!err && (flow->flags & MLX5E_TC_FLOW_ESWITCH)) { rep = rpriv->rep; @@ -2443,10 +2470,21 @@ static int mlx5e_gen_vxlan_header(char buf[], struct ip_tunnel_key *tun_key) } static int mlx5e_gen_ip_tunnel_header(char buf[], __u8 *ip_proto, - struct ip_tunnel_key *tun_key) + struct mlx5e_encap_entry *e) { - *ip_proto = IPPROTO_UDP; - return mlx5e_gen_vxlan_header(buf, tun_key); + int err = 0; + struct ip_tunnel_key *key = &e->tun_info.key; + + if (e->tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN) { + *ip_proto = IPPROTO_UDP; + err = mlx5e_gen_vxlan_header(buf, key); + } else { + pr_warn("mlx5: Cannot generate tunnel header for tunnel type (%d)\n" + , e->tunnel_type); + err = -EOPNOTSUPP; + } + + return err; } static int mlx5e_create_encap_header_ipv4(struct mlx5e_priv *priv, @@ -2456,7 +2494,7 @@ static int mlx5e_create_encap_header_ipv4(struct mlx5e_priv *priv, int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size); int ipv4_encap_size = ETH_HLEN + sizeof(struct iphdr) + - VXLAN_HLEN; + e->tunnel_hlen; struct ip_tunnel_key *tun_key = &e->tun_info.key; struct net_device *out_dev; struct neighbour *n = NULL; @@ -2527,7 +2565,7 @@ static int mlx5e_create_encap_header_ipv4(struct mlx5e_priv *priv, /* add tunneling protocol header */ err = mlx5e_gen_ip_tunnel_header((char *)ip + sizeof(struct iphdr), - &ip->protocol, tun_key); + &ip->protocol, e); if (err) goto destroy_neigh_entry; @@ -2540,7 +2578,8 @@ static int mlx5e_create_encap_header_ipv4(struct mlx5e_priv *priv, goto out; } - err = mlx5_packet_reformat_alloc(priv->mdev, e->tunnel_type, + err = mlx5_packet_reformat_alloc(priv->mdev, + e->reformat_type, ipv4_encap_size, encap_header, MLX5_FLOW_NAMESPACE_FDB, &e->encap_id); @@ -2569,7 +2608,7 @@ static int mlx5e_create_encap_header_ipv6(struct mlx5e_priv *priv, int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size); int ipv6_encap_size = ETH_HLEN + sizeof(struct ipv6hdr) + - VXLAN_HLEN; + e->tunnel_hlen; struct ip_tunnel_key *tun_key = &e->tun_info.key; struct net_device *out_dev; struct neighbour *n = NULL; @@ -2639,7 +2678,7 @@ static int mlx5e_create_encap_header_ipv6(struct mlx5e_priv *priv, /* add tunneling protocol header */ err = mlx5e_gen_ip_tunnel_header((char *)ip6h + sizeof(struct ipv6hdr), - &ip6h->nexthdr, tun_key); + &ip6h->nexthdr, e); if (err) goto destroy_neigh_entry; @@ -2652,7 +2691,8 @@ static int mlx5e_create_encap_header_ipv6(struct mlx5e_priv *priv, goto out; } - err = mlx5_packet_reformat_alloc(priv->mdev, e->tunnel_type, + err = mlx5_packet_reformat_alloc(priv->mdev, + e->reformat_type, ipv6_encap_size, encap_header, MLX5_FLOW_NAMESPACE_FDB, &e->encap_id); @@ -2674,14 +2714,52 @@ out: return err; } +static int mlx5e_get_tunnel_type(struct net_device *tunnel_dev) +{ + if (netif_is_vxlan(tunnel_dev)) + return MLX5E_TC_TUNNEL_TYPE_VXLAN; + else + return MLX5E_TC_TUNNEL_TYPE_UNKNOWN; +} + bool mlx5e_tc_tun_device_to_offload(struct mlx5e_priv *priv, struct net_device *netdev) { - if (netif_is_vxlan(netdev) && + int tunnel_type = mlx5e_get_tunnel_type(netdev); + + if (tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN && MLX5_CAP_ESW(priv->mdev, vxlan_encap_decap)) return true; + else + return false; +} - return false; +static int mlx5e_init_tunnel_attr(struct net_device *tunnel_dev, + struct mlx5e_priv *priv, + struct mlx5e_encap_entry *e, + struct netlink_ext_ack *extack) +{ + e->tunnel_type = mlx5e_get_tunnel_type(tunnel_dev); + + if (e->tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN) { + int dst_port = be16_to_cpu(e->tun_info.key.tp_dst); + + if (!mlx5_vxlan_lookup_port(priv->mdev->vxlan, dst_port)) { + NL_SET_ERR_MSG_MOD(extack, + "vxlan udp dport was not registered with the HW"); + netdev_warn(priv->netdev, + "%d isn't an offloaded vxlan udp dport\n", + dst_port); + return -EOPNOTSUPP; + } + e->reformat_type = MLX5_REFORMAT_TYPE_L2_TO_VXLAN; + e->tunnel_hlen = VXLAN_HLEN; + } else { + e->reformat_type = -1; + e->tunnel_hlen = -1; + return -EOPNOTSUPP; + } + return 0; } static int mlx5e_attach_encap(struct mlx5e_priv *priv, @@ -2696,28 +2774,9 @@ static int mlx5e_attach_encap(struct mlx5e_priv *priv, struct mlx5_esw_flow_attr *attr = flow->esw_attr; struct ip_tunnel_key *key = &tun_info->key; struct mlx5e_encap_entry *e; - int tunnel_type, err = 0; uintptr_t hash_key; bool found = false; - - if (!MLX5_CAP_ESW(priv->mdev, vxlan_encap_decap)) { - NL_SET_ERR_MSG_MOD(extack, - "vxlan HW offloading is not supported"); - netdev_warn(priv->netdev, - "vxlan HW offloading is not supported\n"); - return -EOPNOTSUPP; - } - - if (mlx5_vxlan_lookup_port(priv->mdev->vxlan, be16_to_cpu(key->tp_dst)) && - MLX5_CAP_ESW(priv->mdev, vxlan_encap_decap)) { - tunnel_type = MLX5_REFORMAT_TYPE_L2_TO_VXLAN; - } else { - NL_SET_ERR_MSG_MOD(extack, - "port isn't an offloaded vxlan udp dport"); - netdev_warn(priv->netdev, - "%d isn't an offloaded vxlan udp dport\n", be16_to_cpu(key->tp_dst)); - return -EOPNOTSUPP; - } + int err = 0; hash_key = hash_encap_info(key); @@ -2738,7 +2797,10 @@ static int mlx5e_attach_encap(struct mlx5e_priv *priv, return -ENOMEM; e->tun_info = *tun_info; - e->tunnel_type = tunnel_type; + err = mlx5e_init_tunnel_attr(mirred_dev, priv, e, extack); + if (err) + goto out_err; + INIT_LIST_HEAD(&e->flows); if (family == AF_INET) @@ -3063,7 +3125,8 @@ mlx5e_add_fdb_flow(struct mlx5e_priv *priv, goto out; parse_attr->filter_dev = filter_dev; flow->esw_attr->parse_attr = parse_attr; - err = parse_cls_flower(flow->priv, flow, &parse_attr->spec, f); + err = parse_cls_flower(flow->priv, flow, &parse_attr->spec, + f, filter_dev); if (err) goto err_free; @@ -3116,7 +3179,8 @@ mlx5e_add_nic_flow(struct mlx5e_priv *priv, goto out; parse_attr->filter_dev = filter_dev; - err = parse_cls_flower(flow->priv, flow, &parse_attr->spec, f); + err = parse_cls_flower(flow->priv, flow, &parse_attr->spec, + f, filter_dev); if (err) goto err_free; -- cgit v1.2.3-55-g7522 From 101f4de9dd521c6d06dfdacaa35e506a8db8494b Mon Sep 17 00:00:00 2001 From: Oz Shlomo Date: Sun, 2 Dec 2018 14:43:27 +0200 Subject: net/mlx5e: Move TC tunnel offloading code to separate source file Move tunnel offloading related code to a separate source file for better code maintainability. Code refactoring with no functional change. Signed-off-by: Oz Shlomo Reviewed-by: Eli Britstein Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/Makefile | 2 +- .../net/ethernet/mellanox/mlx5/core/en/tc_tun.c | 496 ++++++++++++++++++++ .../net/ethernet/mellanox/mlx5/core/en/tc_tun.h | 43 ++ drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 1 + drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 500 +-------------------- drivers/net/ethernet/mellanox/mlx5/core/en_tc.h | 2 - 6 files changed, 548 insertions(+), 496 deletions(-) create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile index d499b3d00348..40764d87413a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile +++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile @@ -30,7 +30,7 @@ mlx5_core-$(CONFIG_MLX5_CORE_EN) += en_main.o en_common.o en_fs.o en_ethtool.o \ mlx5_core-$(CONFIG_MLX5_EN_ARFS) += en_arfs.o mlx5_core-$(CONFIG_MLX5_EN_RXNFC) += en_fs_ethtool.o mlx5_core-$(CONFIG_MLX5_CORE_EN_DCB) += en_dcbnl.o en/port_buffer.o -mlx5_core-$(CONFIG_MLX5_ESWITCH) += en_rep.o en_tc.o +mlx5_core-$(CONFIG_MLX5_ESWITCH) += en_rep.o en_tc.o en/tc_tun.o # # Core extra diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c new file mode 100644 index 000000000000..eaa43477e0ea --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c @@ -0,0 +1,496 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* Copyright (c) 2018 Mellanox Technologies. */ + +#include +#include "lib/vxlan.h" +#include "en/tc_tun.h" + +static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv, + struct net_device *mirred_dev, + struct net_device **out_dev, + struct flowi4 *fl4, + struct neighbour **out_n, + u8 *out_ttl) +{ + struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; + struct mlx5e_rep_priv *uplink_rpriv; + struct rtable *rt; + struct neighbour *n = NULL; + +#if IS_ENABLED(CONFIG_INET) + int ret; + + rt = ip_route_output_key(dev_net(mirred_dev), fl4); + ret = PTR_ERR_OR_ZERO(rt); + if (ret) + return ret; +#else + return -EOPNOTSUPP; +#endif + uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH); + /* if the egress device isn't on the same HW e-switch, we use the uplink */ + if (!switchdev_port_same_parent_id(priv->netdev, rt->dst.dev)) + *out_dev = uplink_rpriv->netdev; + else + *out_dev = rt->dst.dev; + + if (!(*out_ttl)) + *out_ttl = ip4_dst_hoplimit(&rt->dst); + n = dst_neigh_lookup(&rt->dst, &fl4->daddr); + ip_rt_put(rt); + if (!n) + return -ENOMEM; + + *out_n = n; + return 0; +} + +static const char *mlx5e_netdev_kind(struct net_device *dev) +{ + if (dev->rtnl_link_ops) + return dev->rtnl_link_ops->kind; + else + return ""; +} + +static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv, + struct net_device *mirred_dev, + struct net_device **out_dev, + struct flowi6 *fl6, + struct neighbour **out_n, + u8 *out_ttl) +{ + struct neighbour *n = NULL; + struct dst_entry *dst; + +#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6) + struct mlx5e_rep_priv *uplink_rpriv; + struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; + int ret; + + ret = ipv6_stub->ipv6_dst_lookup(dev_net(mirred_dev), NULL, &dst, + fl6); + if (ret < 0) + return ret; + + if (!(*out_ttl)) + *out_ttl = ip6_dst_hoplimit(dst); + + uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH); + /* if the egress device isn't on the same HW e-switch, we use the uplink */ + if (!switchdev_port_same_parent_id(priv->netdev, dst->dev)) + *out_dev = uplink_rpriv->netdev; + else + *out_dev = dst->dev; +#else + return -EOPNOTSUPP; +#endif + + n = dst_neigh_lookup(dst, &fl6->daddr); + dst_release(dst); + if (!n) + return -ENOMEM; + + *out_n = n; + return 0; +} + +static int mlx5e_gen_vxlan_header(char buf[], struct ip_tunnel_key *tun_key) +{ + __be32 tun_id = tunnel_id_to_key32(tun_key->tun_id); + struct udphdr *udp = (struct udphdr *)(buf); + struct vxlanhdr *vxh = (struct vxlanhdr *) + ((char *)udp + sizeof(struct udphdr)); + + udp->dest = tun_key->tp_dst; + vxh->vx_flags = VXLAN_HF_VNI; + vxh->vx_vni = vxlan_vni_field(tun_id); + + return 0; +} + +static int mlx5e_gen_ip_tunnel_header(char buf[], __u8 *ip_proto, + struct mlx5e_encap_entry *e) +{ + int err = 0; + struct ip_tunnel_key *key = &e->tun_info.key; + + if (e->tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN) { + *ip_proto = IPPROTO_UDP; + err = mlx5e_gen_vxlan_header(buf, key); + } else { + pr_warn("mlx5: Cannot generate tunnel header for tunnel type (%d)\n" + , e->tunnel_type); + err = -EOPNOTSUPP; + } + + return err; +} + +int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv, + struct net_device *mirred_dev, + struct mlx5e_encap_entry *e) +{ + int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size); + int ipv4_encap_size = ETH_HLEN + + sizeof(struct iphdr) + + e->tunnel_hlen; + struct ip_tunnel_key *tun_key = &e->tun_info.key; + struct net_device *out_dev; + struct neighbour *n = NULL; + struct flowi4 fl4 = {}; + char *encap_header; + struct ethhdr *eth; + u8 nud_state, ttl; + struct iphdr *ip; + int err; + + if (max_encap_size < ipv4_encap_size) { + mlx5_core_warn(priv->mdev, "encap size %d too big, max supported is %d\n", + ipv4_encap_size, max_encap_size); + return -EOPNOTSUPP; + } + + encap_header = kzalloc(ipv4_encap_size, GFP_KERNEL); + if (!encap_header) + return -ENOMEM; + + /* add the IP fields */ + fl4.flowi4_tos = tun_key->tos; + fl4.daddr = tun_key->u.ipv4.dst; + fl4.saddr = tun_key->u.ipv4.src; + ttl = tun_key->ttl; + + err = mlx5e_route_lookup_ipv4(priv, mirred_dev, &out_dev, + &fl4, &n, &ttl); + if (err) + goto free_encap; + + /* used by mlx5e_detach_encap to lookup a neigh hash table + * entry in the neigh hash table when a user deletes a rule + */ + e->m_neigh.dev = n->dev; + e->m_neigh.family = n->ops->family; + memcpy(&e->m_neigh.dst_ip, n->primary_key, n->tbl->key_len); + e->out_dev = out_dev; + + /* It's important to add the neigh to the hash table before checking + * the neigh validity state. So if we'll get a notification, in case the + * neigh changes it's validity state, we would find the relevant neigh + * in the hash. + */ + err = mlx5e_rep_encap_entry_attach(netdev_priv(out_dev), e); + if (err) + goto free_encap; + + read_lock_bh(&n->lock); + nud_state = n->nud_state; + ether_addr_copy(e->h_dest, n->ha); + read_unlock_bh(&n->lock); + + /* add ethernet header */ + eth = (struct ethhdr *)encap_header; + ether_addr_copy(eth->h_dest, e->h_dest); + ether_addr_copy(eth->h_source, out_dev->dev_addr); + eth->h_proto = htons(ETH_P_IP); + + /* add ip header */ + ip = (struct iphdr *)((char *)eth + sizeof(struct ethhdr)); + ip->tos = tun_key->tos; + ip->version = 0x4; + ip->ihl = 0x5; + ip->ttl = ttl; + ip->daddr = fl4.daddr; + ip->saddr = fl4.saddr; + + /* add tunneling protocol header */ + err = mlx5e_gen_ip_tunnel_header((char *)ip + sizeof(struct iphdr), + &ip->protocol, e); + if (err) + goto destroy_neigh_entry; + + e->encap_size = ipv4_encap_size; + e->encap_header = encap_header; + + if (!(nud_state & NUD_VALID)) { + neigh_event_send(n, NULL); + err = -EAGAIN; + 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) + goto destroy_neigh_entry; + + e->flags |= MLX5_ENCAP_ENTRY_VALID; + mlx5e_rep_queue_neigh_stats_work(netdev_priv(out_dev)); + neigh_release(n); + return err; + +destroy_neigh_entry: + mlx5e_rep_encap_entry_detach(netdev_priv(e->out_dev), e); +free_encap: + kfree(encap_header); +out: + if (n) + neigh_release(n); + return err; +} + +int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv, + struct net_device *mirred_dev, + struct mlx5e_encap_entry *e) +{ + int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size); + int ipv6_encap_size = ETH_HLEN + + sizeof(struct ipv6hdr) + + e->tunnel_hlen; + struct ip_tunnel_key *tun_key = &e->tun_info.key; + struct net_device *out_dev; + struct neighbour *n = NULL; + struct flowi6 fl6 = {}; + struct ipv6hdr *ip6h; + char *encap_header; + struct ethhdr *eth; + u8 nud_state, ttl; + int err; + + if (max_encap_size < ipv6_encap_size) { + mlx5_core_warn(priv->mdev, "encap size %d too big, max supported is %d\n", + ipv6_encap_size, max_encap_size); + return -EOPNOTSUPP; + } + + encap_header = kzalloc(ipv6_encap_size, GFP_KERNEL); + if (!encap_header) + return -ENOMEM; + + ttl = tun_key->ttl; + + fl6.flowlabel = ip6_make_flowinfo(RT_TOS(tun_key->tos), tun_key->label); + fl6.daddr = tun_key->u.ipv6.dst; + fl6.saddr = tun_key->u.ipv6.src; + + err = mlx5e_route_lookup_ipv6(priv, mirred_dev, &out_dev, + &fl6, &n, &ttl); + if (err) + goto free_encap; + + /* used by mlx5e_detach_encap to lookup a neigh hash table + * entry in the neigh hash table when a user deletes a rule + */ + e->m_neigh.dev = n->dev; + e->m_neigh.family = n->ops->family; + memcpy(&e->m_neigh.dst_ip, n->primary_key, n->tbl->key_len); + e->out_dev = out_dev; + + /* It's importent to add the neigh to the hash table before checking + * the neigh validity state. So if we'll get a notification, in case the + * neigh changes it's validity state, we would find the relevant neigh + * in the hash. + */ + err = mlx5e_rep_encap_entry_attach(netdev_priv(out_dev), e); + if (err) + goto free_encap; + + read_lock_bh(&n->lock); + nud_state = n->nud_state; + ether_addr_copy(e->h_dest, n->ha); + read_unlock_bh(&n->lock); + + /* add ethernet header */ + eth = (struct ethhdr *)encap_header; + ether_addr_copy(eth->h_dest, e->h_dest); + ether_addr_copy(eth->h_source, out_dev->dev_addr); + eth->h_proto = htons(ETH_P_IPV6); + + /* add ip header */ + ip6h = (struct ipv6hdr *)((char *)eth + sizeof(struct ethhdr)); + ip6_flow_hdr(ip6h, tun_key->tos, 0); + /* the HW fills up ipv6 payload len */ + ip6h->hop_limit = ttl; + ip6h->daddr = fl6.daddr; + ip6h->saddr = fl6.saddr; + + /* add tunneling protocol header */ + err = mlx5e_gen_ip_tunnel_header((char *)ip6h + sizeof(struct ipv6hdr), + &ip6h->nexthdr, e); + if (err) + goto destroy_neigh_entry; + + e->encap_size = ipv6_encap_size; + e->encap_header = encap_header; + + if (!(nud_state & NUD_VALID)) { + neigh_event_send(n, NULL); + err = -EAGAIN; + 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) + goto destroy_neigh_entry; + + e->flags |= MLX5_ENCAP_ENTRY_VALID; + mlx5e_rep_queue_neigh_stats_work(netdev_priv(out_dev)); + neigh_release(n); + return err; + +destroy_neigh_entry: + mlx5e_rep_encap_entry_detach(netdev_priv(e->out_dev), e); +free_encap: + kfree(encap_header); +out: + if (n) + neigh_release(n); + return err; +} + +int mlx5e_tc_tun_get_type(struct net_device *tunnel_dev) +{ + if (netif_is_vxlan(tunnel_dev)) + return MLX5E_TC_TUNNEL_TYPE_VXLAN; + else + return MLX5E_TC_TUNNEL_TYPE_UNKNOWN; +} + +bool mlx5e_tc_tun_device_to_offload(struct mlx5e_priv *priv, + struct net_device *netdev) +{ + int tunnel_type = mlx5e_tc_tun_get_type(netdev); + + if (tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN && + MLX5_CAP_ESW(priv->mdev, vxlan_encap_decap)) + return true; + else + return false; +} + +int mlx5e_tc_tun_init_encap_attr(struct net_device *tunnel_dev, + struct mlx5e_priv *priv, + struct mlx5e_encap_entry *e, + struct netlink_ext_ack *extack) +{ + e->tunnel_type = mlx5e_tc_tun_get_type(tunnel_dev); + + if (e->tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN) { + int dst_port = be16_to_cpu(e->tun_info.key.tp_dst); + + if (!mlx5_vxlan_lookup_port(priv->mdev->vxlan, dst_port)) { + NL_SET_ERR_MSG_MOD(extack, + "vxlan udp dport was not registered with the HW"); + netdev_warn(priv->netdev, + "%d isn't an offloaded vxlan udp dport\n", + dst_port); + return -EOPNOTSUPP; + } + e->reformat_type = MLX5_REFORMAT_TYPE_L2_TO_VXLAN; + e->tunnel_hlen = VXLAN_HLEN; + } else { + e->reformat_type = -1; + e->tunnel_hlen = -1; + return -EOPNOTSUPP; + } + return 0; +} + +static int mlx5e_tc_tun_parse_vxlan(struct mlx5e_priv *priv, + struct mlx5_flow_spec *spec, + struct tc_cls_flower_offload *f, + void *headers_c, + void *headers_v) +{ + struct netlink_ext_ack *extack = f->common.extack; + struct flow_dissector_key_ports *key = + skb_flow_dissector_target(f->dissector, + FLOW_DISSECTOR_KEY_ENC_PORTS, + f->key); + struct flow_dissector_key_ports *mask = + skb_flow_dissector_target(f->dissector, + FLOW_DISSECTOR_KEY_ENC_PORTS, + f->mask); + void *misc_c = MLX5_ADDR_OF(fte_match_param, + spec->match_criteria, + misc_parameters); + void *misc_v = MLX5_ADDR_OF(fte_match_param, + spec->match_value, + misc_parameters); + + /* Full udp dst port must be given */ + if (!dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_PORTS) || + memchr_inv(&mask->dst, 0xff, sizeof(mask->dst))) { + NL_SET_ERR_MSG_MOD(extack, + "VXLAN decap filter must include enc_dst_port condition"); + netdev_warn(priv->netdev, + "VXLAN decap filter must include enc_dst_port condition\n"); + return -EOPNOTSUPP; + } + + /* udp dst port must be knonwn as a VXLAN port */ + if (!mlx5_vxlan_lookup_port(priv->mdev->vxlan, be16_to_cpu(key->dst))) { + NL_SET_ERR_MSG_MOD(extack, + "Matched UDP port is not registered as a VXLAN port"); + netdev_warn(priv->netdev, + "UDP port %d is not registered as a VXLAN port\n", + be16_to_cpu(key->dst)); + return -EOPNOTSUPP; + } + + /* dst UDP port is valid here */ + MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, headers_c, ip_protocol); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_protocol, IPPROTO_UDP); + + MLX5_SET(fte_match_set_lyr_2_4, headers_c, udp_dport, ntohs(mask->dst)); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, udp_dport, ntohs(key->dst)); + + MLX5_SET(fte_match_set_lyr_2_4, headers_c, udp_sport, ntohs(mask->src)); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, udp_sport, ntohs(key->src)); + + /* match on VNI */ + if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_KEYID)) { + struct flow_dissector_key_keyid *key = + skb_flow_dissector_target(f->dissector, + FLOW_DISSECTOR_KEY_ENC_KEYID, + f->key); + struct flow_dissector_key_keyid *mask = + skb_flow_dissector_target(f->dissector, + FLOW_DISSECTOR_KEY_ENC_KEYID, + f->mask); + MLX5_SET(fte_match_set_misc, misc_c, vxlan_vni, + be32_to_cpu(mask->keyid)); + MLX5_SET(fte_match_set_misc, misc_v, vxlan_vni, + be32_to_cpu(key->keyid)); + } + return 0; +} + +int mlx5e_tc_tun_parse(struct net_device *filter_dev, + struct mlx5e_priv *priv, + struct mlx5_flow_spec *spec, + struct tc_cls_flower_offload *f, + void *headers_c, + void *headers_v) +{ + int tunnel_type; + int err = 0; + + tunnel_type = mlx5e_tc_tun_get_type(filter_dev); + if (tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN) { + err = mlx5e_tc_tun_parse_vxlan(priv, spec, f, + headers_c, headers_v); + } else { + netdev_warn(priv->netdev, + "decapsulation offload is not supported for %s net device (%d)\n", + mlx5e_netdev_kind(filter_dev), tunnel_type); + return -EOPNOTSUPP; + } + return err; +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h new file mode 100644 index 000000000000..ad4fc93b17a1 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h @@ -0,0 +1,43 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* Copyright (c) 2018 Mellanox Technologies. */ + +#ifndef __MLX5_EN_TC_TUNNEL_H__ +#define __MLX5_EN_TC_TUNNEL_H__ + +#include +#include +#include +#include +#include "en.h" +#include "en_rep.h" + +enum { + MLX5E_TC_TUNNEL_TYPE_UNKNOWN, + MLX5E_TC_TUNNEL_TYPE_VXLAN +}; + +int mlx5e_tc_tun_init_encap_attr(struct net_device *tunnel_dev, + struct mlx5e_priv *priv, + struct mlx5e_encap_entry *e, + struct netlink_ext_ack *extack); + +int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv, + struct net_device *mirred_dev, + struct mlx5e_encap_entry *e); + +int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv, + struct net_device *mirred_dev, + struct mlx5e_encap_entry *e); + +int mlx5e_tc_tun_get_type(struct net_device *tunnel_dev); +bool mlx5e_tc_tun_device_to_offload(struct mlx5e_priv *priv, + struct net_device *netdev); + +int mlx5e_tc_tun_parse(struct net_device *filter_dev, + struct mlx5e_priv *priv, + struct mlx5_flow_spec *spec, + struct tc_cls_flower_offload *f, + void *headers_c, + void *headers_v); + +#endif //__MLX5_EN_TC_TUNNEL_H__ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index 5cac4de435c9..85e51bd4147f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -42,6 +42,7 @@ #include "en.h" #include "en_rep.h" #include "en_tc.h" +#include "en/tc_tun.h" #include "fs_core.h" #define MLX5E_REP_PARAMS_LOG_SQ_SIZE \ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 7d7f490d8ff1..abc200947e84 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -44,15 +44,14 @@ #include #include #include -#include #include #include "en.h" #include "en_rep.h" #include "en_tc.h" #include "eswitch.h" -#include "lib/vxlan.h" #include "fs_core.h" #include "en/port.h" +#include "en/tc_tun.h" struct mlx5_nic_flow_attr { u32 action; @@ -79,13 +78,6 @@ enum { #define MLX5E_TC_MAX_SPLITS 1 -enum { - MLX5E_TC_TUNNEL_TYPE_UNKNOWN, - MLX5E_TC_TUNNEL_TYPE_VXLAN -}; - -static int mlx5e_get_tunnel_type(struct net_device *tunnel_dev); - struct mlx5e_tc_flow { struct rhash_head node; struct mlx5e_priv *priv; @@ -684,14 +676,6 @@ static void mlx5e_hairpin_flow_del(struct mlx5e_priv *priv, } } -static const char *mlx5e_netdev_kind(struct net_device *dev) -{ - if (dev->rtnl_link_ops) - return dev->rtnl_link_ops->kind; - else - return ""; -} - static int mlx5e_tc_add_nic_flow(struct mlx5e_priv *priv, struct mlx5e_tc_flow_parse_attr *parse_attr, @@ -1209,75 +1193,6 @@ static void mlx5e_tc_del_flow(struct mlx5e_priv *priv, mlx5e_tc_del_nic_flow(priv, flow); } -static int parse_tunnel_vxlan_attr(struct mlx5e_priv *priv, - struct mlx5_flow_spec *spec, - struct tc_cls_flower_offload *f, - void *headers_c, - void *headers_v) -{ - struct netlink_ext_ack *extack = f->common.extack; - struct flow_dissector_key_ports *key = - skb_flow_dissector_target(f->dissector, - FLOW_DISSECTOR_KEY_ENC_PORTS, - f->key); - struct flow_dissector_key_ports *mask = - skb_flow_dissector_target(f->dissector, - FLOW_DISSECTOR_KEY_ENC_PORTS, - f->mask); - void *misc_c = MLX5_ADDR_OF(fte_match_param, - spec->match_criteria, - misc_parameters); - void *misc_v = MLX5_ADDR_OF(fte_match_param, - spec->match_value, - misc_parameters); - - /* Full udp dst port must be given */ - if (!dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_PORTS) || - memchr_inv(&mask->dst, 0xff, sizeof(mask->dst))) { - NL_SET_ERR_MSG_MOD(extack, - "VXLAN decap filter must include enc_dst_port condition"); - netdev_warn(priv->netdev, - "VXLAN decap filter must include enc_dst_port condition\n"); - return -EOPNOTSUPP; - } - - /* udp dst port must be knonwn as a VXLAN port */ - if (!mlx5_vxlan_lookup_port(priv->mdev->vxlan, be16_to_cpu(key->dst))) { - NL_SET_ERR_MSG_MOD(extack, - "Matched UDP port is not registered as a VXLAN port"); - netdev_warn(priv->netdev, - "UDP port %d is not registered as a VXLAN port\n", - be16_to_cpu(key->dst)); - return -EOPNOTSUPP; - } - - /* dst UDP port is valid here */ - MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, headers_c, ip_protocol); - MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_protocol, IPPROTO_UDP); - - MLX5_SET(fte_match_set_lyr_2_4, headers_c, udp_dport, ntohs(mask->dst)); - MLX5_SET(fte_match_set_lyr_2_4, headers_v, udp_dport, ntohs(key->dst)); - - MLX5_SET(fte_match_set_lyr_2_4, headers_c, udp_sport, ntohs(mask->src)); - MLX5_SET(fte_match_set_lyr_2_4, headers_v, udp_sport, ntohs(key->src)); - - /* match on VNI */ - if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_KEYID)) { - struct flow_dissector_key_keyid *key = - skb_flow_dissector_target(f->dissector, - FLOW_DISSECTOR_KEY_ENC_KEYID, - f->key); - struct flow_dissector_key_keyid *mask = - skb_flow_dissector_target(f->dissector, - FLOW_DISSECTOR_KEY_ENC_KEYID, - f->mask); - MLX5_SET(fte_match_set_misc, misc_c, vxlan_vni, - be32_to_cpu(mask->keyid)); - MLX5_SET(fte_match_set_misc, misc_v, vxlan_vni, - be32_to_cpu(key->keyid)); - } - return 0; -} static int parse_tunnel_attr(struct mlx5e_priv *priv, struct mlx5_flow_spec *spec, @@ -1294,29 +1209,14 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv, skb_flow_dissector_target(f->dissector, FLOW_DISSECTOR_KEY_ENC_CONTROL, f->key); - int tunnel_type; int err = 0; - tunnel_type = mlx5e_get_tunnel_type(filter_dev); - if (tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN) { - err = parse_tunnel_vxlan_attr(priv, spec, f, - headers_c, headers_v); - } else { - NL_SET_ERR_MSG_MOD(extack, - "decapsulation offload is not supported"); - netdev_warn(priv->netdev, - "decapsulation offload is not supported for %s net device (%d)\n", - mlx5e_netdev_kind(filter_dev), tunnel_type); - return -EOPNOTSUPP; - } - + err = mlx5e_tc_tun_parse(filter_dev, priv, spec, f, + headers_c, headers_v); if (err) { NL_SET_ERR_MSG_MOD(extack, "failed to parse tunnel attributes"); - netdev_warn(priv->netdev, - "failed to parse %s tunnel attributes (%d)\n", - mlx5e_netdev_kind(filter_dev), tunnel_type); - return -EOPNOTSUPP; + return err; } if (enc_control->addr_type == FLOW_DISSECTOR_KEY_IPV4_ADDRS) { @@ -2359,45 +2259,6 @@ static inline int hash_encap_info(struct ip_tunnel_key *key) return jhash(key, sizeof(*key), 0); } -static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv, - struct net_device *mirred_dev, - struct net_device **out_dev, - struct flowi4 *fl4, - struct neighbour **out_n, - u8 *out_ttl) -{ - struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; - struct mlx5e_rep_priv *uplink_rpriv; - struct rtable *rt; - struct neighbour *n = NULL; - -#if IS_ENABLED(CONFIG_INET) - int ret; - - rt = ip_route_output_key(dev_net(mirred_dev), fl4); - ret = PTR_ERR_OR_ZERO(rt); - if (ret) - return ret; -#else - return -EOPNOTSUPP; -#endif - uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH); - /* if the egress device isn't on the same HW e-switch, we use the uplink */ - if (!switchdev_port_same_parent_id(priv->netdev, rt->dst.dev)) - *out_dev = uplink_rpriv->netdev; - else - *out_dev = rt->dst.dev; - - if (!(*out_ttl)) - *out_ttl = ip4_dst_hoplimit(&rt->dst); - n = dst_neigh_lookup(&rt->dst, &fl4->daddr); - ip_rt_put(rt); - if (!n) - return -ENOMEM; - - *out_n = n; - return 0; -} static bool is_merged_eswitch_dev(struct mlx5e_priv *priv, struct net_device *peer_netdev) @@ -2413,354 +2274,7 @@ static bool is_merged_eswitch_dev(struct mlx5e_priv *priv, (peer_priv->mdev->priv.eswitch->mode == SRIOV_OFFLOADS)); } -static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv, - struct net_device *mirred_dev, - struct net_device **out_dev, - struct flowi6 *fl6, - struct neighbour **out_n, - u8 *out_ttl) -{ - struct neighbour *n = NULL; - struct dst_entry *dst; - -#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6) - struct mlx5e_rep_priv *uplink_rpriv; - struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; - int ret; - - ret = ipv6_stub->ipv6_dst_lookup(dev_net(mirred_dev), NULL, &dst, - fl6); - if (ret < 0) - return ret; - - if (!(*out_ttl)) - *out_ttl = ip6_dst_hoplimit(dst); - - uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH); - /* if the egress device isn't on the same HW e-switch, we use the uplink */ - if (!switchdev_port_same_parent_id(priv->netdev, dst->dev)) - *out_dev = uplink_rpriv->netdev; - else - *out_dev = dst->dev; -#else - return -EOPNOTSUPP; -#endif - - n = dst_neigh_lookup(dst, &fl6->daddr); - dst_release(dst); - if (!n) - return -ENOMEM; - - *out_n = n; - return 0; -} - -static int mlx5e_gen_vxlan_header(char buf[], struct ip_tunnel_key *tun_key) -{ - __be32 tun_id = tunnel_id_to_key32(tun_key->tun_id); - struct udphdr *udp = (struct udphdr *)(buf); - struct vxlanhdr *vxh = (struct vxlanhdr *) - ((char *)udp + sizeof(struct udphdr)); - - udp->dest = tun_key->tp_dst; - vxh->vx_flags = VXLAN_HF_VNI; - vxh->vx_vni = vxlan_vni_field(tun_id); - - return 0; -} - -static int mlx5e_gen_ip_tunnel_header(char buf[], __u8 *ip_proto, - struct mlx5e_encap_entry *e) -{ - int err = 0; - struct ip_tunnel_key *key = &e->tun_info.key; - - if (e->tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN) { - *ip_proto = IPPROTO_UDP; - err = mlx5e_gen_vxlan_header(buf, key); - } else { - pr_warn("mlx5: Cannot generate tunnel header for tunnel type (%d)\n" - , e->tunnel_type); - err = -EOPNOTSUPP; - } - - return err; -} - -static int mlx5e_create_encap_header_ipv4(struct mlx5e_priv *priv, - struct net_device *mirred_dev, - struct mlx5e_encap_entry *e) -{ - int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size); - int ipv4_encap_size = ETH_HLEN + - sizeof(struct iphdr) + - e->tunnel_hlen; - struct ip_tunnel_key *tun_key = &e->tun_info.key; - struct net_device *out_dev; - struct neighbour *n = NULL; - struct flowi4 fl4 = {}; - char *encap_header; - struct ethhdr *eth; - u8 nud_state, ttl; - struct iphdr *ip; - int err; - - if (max_encap_size < ipv4_encap_size) { - mlx5_core_warn(priv->mdev, "encap size %d too big, max supported is %d\n", - ipv4_encap_size, max_encap_size); - return -EOPNOTSUPP; - } - - encap_header = kzalloc(ipv4_encap_size, GFP_KERNEL); - if (!encap_header) - return -ENOMEM; - - /* add the IP fields */ - fl4.flowi4_tos = tun_key->tos; - fl4.daddr = tun_key->u.ipv4.dst; - fl4.saddr = tun_key->u.ipv4.src; - ttl = tun_key->ttl; - - err = mlx5e_route_lookup_ipv4(priv, mirred_dev, &out_dev, - &fl4, &n, &ttl); - if (err) - goto free_encap; - - /* used by mlx5e_detach_encap to lookup a neigh hash table - * entry in the neigh hash table when a user deletes a rule - */ - e->m_neigh.dev = n->dev; - e->m_neigh.family = n->ops->family; - memcpy(&e->m_neigh.dst_ip, n->primary_key, n->tbl->key_len); - e->out_dev = out_dev; - - /* It's important to add the neigh to the hash table before checking - * the neigh validity state. So if we'll get a notification, in case the - * neigh changes it's validity state, we would find the relevant neigh - * in the hash. - */ - err = mlx5e_rep_encap_entry_attach(netdev_priv(out_dev), e); - if (err) - goto free_encap; - - read_lock_bh(&n->lock); - nud_state = n->nud_state; - ether_addr_copy(e->h_dest, n->ha); - read_unlock_bh(&n->lock); - - /* add ethernet header */ - eth = (struct ethhdr *)encap_header; - ether_addr_copy(eth->h_dest, e->h_dest); - ether_addr_copy(eth->h_source, out_dev->dev_addr); - eth->h_proto = htons(ETH_P_IP); - - /* add ip header */ - ip = (struct iphdr *)((char *)eth + sizeof(struct ethhdr)); - ip->tos = tun_key->tos; - ip->version = 0x4; - ip->ihl = 0x5; - ip->ttl = ttl; - ip->daddr = fl4.daddr; - ip->saddr = fl4.saddr; - - /* add tunneling protocol header */ - err = mlx5e_gen_ip_tunnel_header((char *)ip + sizeof(struct iphdr), - &ip->protocol, e); - if (err) - goto destroy_neigh_entry; - - e->encap_size = ipv4_encap_size; - e->encap_header = encap_header; - - if (!(nud_state & NUD_VALID)) { - neigh_event_send(n, NULL); - err = -EAGAIN; - 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) - goto destroy_neigh_entry; - - e->flags |= MLX5_ENCAP_ENTRY_VALID; - mlx5e_rep_queue_neigh_stats_work(netdev_priv(out_dev)); - neigh_release(n); - return err; - -destroy_neigh_entry: - mlx5e_rep_encap_entry_detach(netdev_priv(e->out_dev), e); -free_encap: - kfree(encap_header); -out: - if (n) - neigh_release(n); - return err; -} - -static int mlx5e_create_encap_header_ipv6(struct mlx5e_priv *priv, - struct net_device *mirred_dev, - struct mlx5e_encap_entry *e) -{ - int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size); - int ipv6_encap_size = ETH_HLEN + - sizeof(struct ipv6hdr) + - e->tunnel_hlen; - struct ip_tunnel_key *tun_key = &e->tun_info.key; - struct net_device *out_dev; - struct neighbour *n = NULL; - struct flowi6 fl6 = {}; - struct ipv6hdr *ip6h; - char *encap_header; - struct ethhdr *eth; - u8 nud_state, ttl; - int err; - - if (max_encap_size < ipv6_encap_size) { - mlx5_core_warn(priv->mdev, "encap size %d too big, max supported is %d\n", - ipv6_encap_size, max_encap_size); - return -EOPNOTSUPP; - } - - encap_header = kzalloc(ipv6_encap_size, GFP_KERNEL); - if (!encap_header) - return -ENOMEM; - - ttl = tun_key->ttl; - - fl6.flowlabel = ip6_make_flowinfo(RT_TOS(tun_key->tos), tun_key->label); - fl6.daddr = tun_key->u.ipv6.dst; - fl6.saddr = tun_key->u.ipv6.src; - - err = mlx5e_route_lookup_ipv6(priv, mirred_dev, &out_dev, - &fl6, &n, &ttl); - if (err) - goto free_encap; - - /* used by mlx5e_detach_encap to lookup a neigh hash table - * entry in the neigh hash table when a user deletes a rule - */ - e->m_neigh.dev = n->dev; - e->m_neigh.family = n->ops->family; - memcpy(&e->m_neigh.dst_ip, n->primary_key, n->tbl->key_len); - e->out_dev = out_dev; - - /* It's importent to add the neigh to the hash table before checking - * the neigh validity state. So if we'll get a notification, in case the - * neigh changes it's validity state, we would find the relevant neigh - * in the hash. - */ - err = mlx5e_rep_encap_entry_attach(netdev_priv(out_dev), e); - if (err) - goto free_encap; - - read_lock_bh(&n->lock); - nud_state = n->nud_state; - ether_addr_copy(e->h_dest, n->ha); - read_unlock_bh(&n->lock); - - /* add ethernet header */ - eth = (struct ethhdr *)encap_header; - ether_addr_copy(eth->h_dest, e->h_dest); - ether_addr_copy(eth->h_source, out_dev->dev_addr); - eth->h_proto = htons(ETH_P_IPV6); - - /* add ip header */ - ip6h = (struct ipv6hdr *)((char *)eth + sizeof(struct ethhdr)); - ip6_flow_hdr(ip6h, tun_key->tos, 0); - /* the HW fills up ipv6 payload len */ - ip6h->hop_limit = ttl; - ip6h->daddr = fl6.daddr; - ip6h->saddr = fl6.saddr; - - /* add tunneling protocol header */ - err = mlx5e_gen_ip_tunnel_header((char *)ip6h + sizeof(struct ipv6hdr), - &ip6h->nexthdr, e); - if (err) - goto destroy_neigh_entry; - - e->encap_size = ipv6_encap_size; - e->encap_header = encap_header; - - if (!(nud_state & NUD_VALID)) { - neigh_event_send(n, NULL); - err = -EAGAIN; - 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) - goto destroy_neigh_entry; - - e->flags |= MLX5_ENCAP_ENTRY_VALID; - mlx5e_rep_queue_neigh_stats_work(netdev_priv(out_dev)); - neigh_release(n); - return err; - -destroy_neigh_entry: - mlx5e_rep_encap_entry_detach(netdev_priv(e->out_dev), e); -free_encap: - kfree(encap_header); -out: - if (n) - neigh_release(n); - return err; -} - -static int mlx5e_get_tunnel_type(struct net_device *tunnel_dev) -{ - if (netif_is_vxlan(tunnel_dev)) - return MLX5E_TC_TUNNEL_TYPE_VXLAN; - else - return MLX5E_TC_TUNNEL_TYPE_UNKNOWN; -} - -bool mlx5e_tc_tun_device_to_offload(struct mlx5e_priv *priv, - struct net_device *netdev) -{ - int tunnel_type = mlx5e_get_tunnel_type(netdev); - - if (tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN && - MLX5_CAP_ESW(priv->mdev, vxlan_encap_decap)) - return true; - else - return false; -} -static int mlx5e_init_tunnel_attr(struct net_device *tunnel_dev, - struct mlx5e_priv *priv, - struct mlx5e_encap_entry *e, - struct netlink_ext_ack *extack) -{ - e->tunnel_type = mlx5e_get_tunnel_type(tunnel_dev); - - if (e->tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN) { - int dst_port = be16_to_cpu(e->tun_info.key.tp_dst); - - if (!mlx5_vxlan_lookup_port(priv->mdev->vxlan, dst_port)) { - NL_SET_ERR_MSG_MOD(extack, - "vxlan udp dport was not registered with the HW"); - netdev_warn(priv->netdev, - "%d isn't an offloaded vxlan udp dport\n", - dst_port); - return -EOPNOTSUPP; - } - e->reformat_type = MLX5_REFORMAT_TYPE_L2_TO_VXLAN; - e->tunnel_hlen = VXLAN_HLEN; - } else { - e->reformat_type = -1; - e->tunnel_hlen = -1; - return -EOPNOTSUPP; - } - return 0; -} static int mlx5e_attach_encap(struct mlx5e_priv *priv, struct ip_tunnel_info *tun_info, @@ -2797,16 +2311,16 @@ static int mlx5e_attach_encap(struct mlx5e_priv *priv, return -ENOMEM; e->tun_info = *tun_info; - err = mlx5e_init_tunnel_attr(mirred_dev, priv, e, extack); + err = mlx5e_tc_tun_init_encap_attr(mirred_dev, priv, e, extack); if (err) goto out_err; INIT_LIST_HEAD(&e->flows); if (family == AF_INET) - err = mlx5e_create_encap_header_ipv4(priv, mirred_dev, e); + err = mlx5e_tc_tun_create_header_ipv4(priv, mirred_dev, e); else if (family == AF_INET6) - err = mlx5e_create_encap_header_ipv6(priv, mirred_dev, e); + err = mlx5e_tc_tun_create_header_ipv6(priv, mirred_dev, e); if (err && err != -EAGAIN) goto out_err; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h index 004e679a4f53..a15c08a35054 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h @@ -70,8 +70,6 @@ void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe); int mlx5e_tc_num_filters(struct mlx5e_priv *priv); -bool mlx5e_tc_tun_device_to_offload(struct mlx5e_priv *priv, - struct net_device *netdev); #else /* CONFIG_MLX5_ESWITCH */ static inline int mlx5e_tc_nic_init(struct mlx5e_priv *priv) { return 0; } -- cgit v1.2.3-55-g7522 From df2ef3bff193229973830fd3fd8acf29fa92715e Mon Sep 17 00:00:00 2001 From: Oz Shlomo Date: Mon, 29 Oct 2018 08:54:42 +0200 Subject: net/mlx5e: Add GRE protocol offloading Add HW offloading support for TC flower filters configured on gretap/ip6gretap net devices. Signed-off-by: Oz Shlomo Reviewed-by: Eli Britstein Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/en/tc_tun.c | 87 ++++++++++++++++++++++ .../net/ethernet/mellanox/mlx5/core/en/tc_tun.h | 3 +- 2 files changed, 89 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') 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 eaa43477e0ea..c1515f013501 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c @@ -2,6 +2,7 @@ /* Copyright (c) 2018 Mellanox Technologies. */ #include +#include #include "lib/vxlan.h" #include "en/tc_tun.h" @@ -109,6 +110,30 @@ static int mlx5e_gen_vxlan_header(char buf[], struct ip_tunnel_key *tun_key) return 0; } +static int mlx5e_gen_gre_header(char buf[], struct ip_tunnel_key *tun_key) +{ + __be32 tun_id = tunnel_id_to_key32(tun_key->tun_id); + int hdr_len; + struct gre_base_hdr *greh = (struct gre_base_hdr *)(buf); + + /* the HW does not calculate GRE csum or sequences */ + if (tun_key->tun_flags & (TUNNEL_CSUM | TUNNEL_SEQ)) + return -EOPNOTSUPP; + + greh->protocol = htons(ETH_P_TEB); + + /* GRE key */ + hdr_len = gre_calc_hlen(tun_key->tun_flags); + greh->flags = gre_tnl_flags_to_gre_flags(tun_key->tun_flags); + if (tun_key->tun_flags & TUNNEL_KEY) { + __be32 *ptr = (__be32 *)(((u8 *)greh) + hdr_len - 4); + + *ptr = tun_id; + } + + return 0; +} + static int mlx5e_gen_ip_tunnel_header(char buf[], __u8 *ip_proto, struct mlx5e_encap_entry *e) { @@ -118,6 +143,9 @@ static int mlx5e_gen_ip_tunnel_header(char buf[], __u8 *ip_proto, if (e->tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN) { *ip_proto = IPPROTO_UDP; err = mlx5e_gen_vxlan_header(buf, key); + } else if (e->tunnel_type == MLX5E_TC_TUNNEL_TYPE_GRETAP) { + *ip_proto = IPPROTO_GRE; + err = mlx5e_gen_gre_header(buf, key); } else { pr_warn("mlx5: Cannot generate tunnel header for tunnel type (%d)\n" , e->tunnel_type); @@ -358,6 +386,9 @@ int mlx5e_tc_tun_get_type(struct net_device *tunnel_dev) { if (netif_is_vxlan(tunnel_dev)) return MLX5E_TC_TUNNEL_TYPE_VXLAN; + else if (netif_is_gretap(tunnel_dev) || + netif_is_ip6gretap(tunnel_dev)) + return MLX5E_TC_TUNNEL_TYPE_GRETAP; else return MLX5E_TC_TUNNEL_TYPE_UNKNOWN; } @@ -370,6 +401,9 @@ bool mlx5e_tc_tun_device_to_offload(struct mlx5e_priv *priv, if (tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN && MLX5_CAP_ESW(priv->mdev, vxlan_encap_decap)) return true; + else if (tunnel_type == MLX5E_TC_TUNNEL_TYPE_GRETAP && + MLX5_CAP_ESW(priv->mdev, nvgre_encap_decap)) + return true; else return false; } @@ -394,6 +428,9 @@ int mlx5e_tc_tun_init_encap_attr(struct net_device *tunnel_dev, } e->reformat_type = MLX5_REFORMAT_TYPE_L2_TO_VXLAN; e->tunnel_hlen = VXLAN_HLEN; + } else if (e->tunnel_type == MLX5E_TC_TUNNEL_TYPE_GRETAP) { + e->reformat_type = MLX5_REFORMAT_TYPE_L2_TO_NVGRE; + e->tunnel_hlen = gre_calc_hlen(e->tun_info.key.tun_flags); } else { e->reformat_type = -1; e->tunnel_hlen = -1; @@ -472,6 +509,53 @@ static int mlx5e_tc_tun_parse_vxlan(struct mlx5e_priv *priv, return 0; } +static int mlx5e_tc_tun_parse_gretap(struct mlx5e_priv *priv, + struct mlx5_flow_spec *spec, + struct tc_cls_flower_offload *f, + void *outer_headers_c, + void *outer_headers_v) +{ + void *misc_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, + misc_parameters); + void *misc_v = MLX5_ADDR_OF(fte_match_param, spec->match_value, + misc_parameters); + + if (!MLX5_CAP_ESW(priv->mdev, nvgre_encap_decap)) { + NL_SET_ERR_MSG_MOD(f->common.extack, + "GRE HW offloading is not supported"); + netdev_warn(priv->netdev, "GRE HW offloading is not supported\n"); + return -EOPNOTSUPP; + } + + MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c, ip_protocol); + MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v, + ip_protocol, IPPROTO_GRE); + + /* gre protocol*/ + MLX5_SET_TO_ONES(fte_match_set_misc, misc_c, gre_protocol); + MLX5_SET(fte_match_set_misc, misc_v, gre_protocol, ETH_P_TEB); + + /* gre key */ + if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_KEYID)) { + struct flow_dissector_key_keyid *mask = NULL; + struct flow_dissector_key_keyid *key = NULL; + + mask = skb_flow_dissector_target(f->dissector, + FLOW_DISSECTOR_KEY_ENC_KEYID, + f->mask); + MLX5_SET(fte_match_set_misc, misc_c, + gre_key.key, be32_to_cpu(mask->keyid)); + + key = skb_flow_dissector_target(f->dissector, + FLOW_DISSECTOR_KEY_ENC_KEYID, + f->key); + MLX5_SET(fte_match_set_misc, misc_v, + gre_key.key, be32_to_cpu(key->keyid)); + } + + return 0; +} + int mlx5e_tc_tun_parse(struct net_device *filter_dev, struct mlx5e_priv *priv, struct mlx5_flow_spec *spec, @@ -486,6 +570,9 @@ int mlx5e_tc_tun_parse(struct net_device *filter_dev, if (tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN) { err = mlx5e_tc_tun_parse_vxlan(priv, spec, f, headers_c, headers_v); + } else if (tunnel_type == MLX5E_TC_TUNNEL_TYPE_GRETAP) { + err = mlx5e_tc_tun_parse_gretap(priv, spec, f, + headers_c, headers_v); } else { netdev_warn(priv->netdev, "decapsulation offload is not supported for %s net device (%d)\n", diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h index ad4fc93b17a1..706ce7bf15e7 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h @@ -13,7 +13,8 @@ enum { MLX5E_TC_TUNNEL_TYPE_UNKNOWN, - MLX5E_TC_TUNNEL_TYPE_VXLAN + MLX5E_TC_TUNNEL_TYPE_VXLAN, + MLX5E_TC_TUNNEL_TYPE_GRETAP }; int mlx5e_tc_tun_init_encap_attr(struct net_device *tunnel_dev, -- cgit v1.2.3-55-g7522 From 1228e912c9343d35b21663af82dea39dd8a575b8 Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Sun, 18 Nov 2018 08:25:17 +0200 Subject: net/mlx5: Consider encapsulation properties when comparing destinations Currently the driver identifies identical vport destinations by comparing the vport ID. The FW extended destination feature enables the driver to forward the packet to the same vport with multiple encapsulation properties. Change the vport destination comparison logic to compare the encapsulation properties in addition to the vport ID. Signed-off-by: Eli Britstein Acked-by: Or Gerlitz Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c index 9d73eb955f75..f21277e636a3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c @@ -1373,7 +1373,10 @@ static bool mlx5_flow_dests_cmp(struct mlx5_flow_destination *d1, { if (d1->type == d2->type) { if ((d1->type == MLX5_FLOW_DESTINATION_TYPE_VPORT && - d1->vport.num == d2->vport.num) || + d1->vport.num == d2->vport.num && + d1->vport.flags == d2->vport.flags && + ((d1->vport.flags & MLX5_FLOW_DEST_VPORT_REFORMAT_ID) ? + (d1->vport.reformat_id == d2->vport.reformat_id) : true)) || (d1->type == MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE && d1->ft == d2->ft) || (d1->type == MLX5_FLOW_DESTINATION_TYPE_TIR && -- cgit v1.2.3-55-g7522 From e85e02bad29e822c3ebde80e7c9e7428329c4a5b Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Fri, 23 Nov 2018 09:38:27 +0200 Subject: net/mlx5: E-Switch, Rename esw attr mirror count field The mirror count esw attributes field is used to determine if splitting the rule to two FTEs is required while programming e-switch mirroring. Rename it to split count, making it clearer with no functional change. Signed-off-by: Eli Britstein Reviewed-by: Or Gerlitz Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 14 +++++++------- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 2 +- drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 14 +++++++------- 3 files changed, 15 insertions(+), 15 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index abc200947e84..342f807543a0 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -838,7 +838,7 @@ mlx5e_tc_offload_fdb_rules(struct mlx5_eswitch *esw, if (IS_ERR(rule)) return rule; - if (attr->mirror_count) { + if (attr->split_count) { flow->rule[1] = mlx5_eswitch_add_fwd_rule(esw, spec, attr); if (IS_ERR(flow->rule[1])) { mlx5_eswitch_del_offloaded_rule(esw, rule, attr); @@ -857,7 +857,7 @@ mlx5e_tc_unoffload_fdb_rules(struct mlx5_eswitch *esw, { flow->flags &= ~MLX5E_TC_FLOW_OFFLOADED; - if (attr->mirror_count) + if (attr->split_count) mlx5_eswitch_del_fwd_rule(esw, flow->rule[1], attr); mlx5_eswitch_del_offloaded_rule(esw, flow->rule[0], attr); @@ -873,7 +873,7 @@ mlx5e_tc_offload_to_slow_path(struct mlx5_eswitch *esw, memcpy(slow_attr, flow->esw_attr, sizeof(*slow_attr)); slow_attr->action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST, - slow_attr->mirror_count = 0, + slow_attr->split_count = 0, slow_attr->dest_chain = FDB_SLOW_PATH_CHAIN, rule = mlx5e_tc_offload_fdb_rules(esw, flow, spec, slow_attr); @@ -2427,7 +2427,7 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, return err; action |= MLX5_FLOW_CONTEXT_ACTION_MOD_HDR; - attr->mirror_count = attr->out_count; + attr->split_count = attr->out_count; continue; } @@ -2501,7 +2501,7 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, encap = true; else return -EOPNOTSUPP; - attr->mirror_count = attr->out_count; + attr->split_count = attr->out_count; continue; } @@ -2511,7 +2511,7 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, if (err) return err; - attr->mirror_count = attr->out_count; + attr->split_count = attr->out_count; continue; } @@ -2546,7 +2546,7 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, if (!actions_match_supported(priv, exts, parse_attr, flow, extack)) return -EOPNOTSUPP; - if (attr->mirror_count > 0 && !mlx5_esw_has_fwd_fdb(priv->mdev)) { + if (attr->split_count > 0 && !mlx5_esw_has_fwd_fdb(priv->mdev)) { NL_SET_ERR_MSG_MOD(extack, "current firmware doesn't support split rule for port mirroring"); netdev_warn_once(priv->netdev, "current firmware doesn't support split rule for port mirroring\n"); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h index 480ffa294867..c28ff6487dfd 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h @@ -287,7 +287,7 @@ struct mlx5_esw_flow_attr { struct mlx5_core_dev *out_mdev[MLX5_MAX_FLOW_FWD_VPORTS]; struct mlx5_core_dev *in_mdev; - int mirror_count; + int split_count; int out_count; int action; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 4d7b65df32ef..63ac9d3d1690 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -81,7 +81,7 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw, { struct mlx5_flow_destination dest[MLX5_MAX_FLOW_FWD_VPORTS + 1] = {}; struct mlx5_flow_act flow_act = { .flags = FLOW_ACT_NO_APPEND, }; - bool mirror = !!(attr->mirror_count); + bool split = !!(attr->split_count); struct mlx5_flow_handle *rule; struct mlx5_flow_table *fdb; int j, i = 0; @@ -120,7 +120,7 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw, dest[i].ft = ft; i++; } else { - for (j = attr->mirror_count; j < attr->out_count; j++) { + for (j = attr->split_count; j < attr->out_count; j++) { dest[i].type = MLX5_FLOW_DESTINATION_TYPE_VPORT; dest[i].vport.num = attr->out_rep[j]->vport; dest[i].vport.vhca_id = @@ -167,7 +167,7 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw, if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT) flow_act.reformat_id = attr->encap_id; - fdb = esw_get_prio_table(esw, attr->chain, attr->prio, !!mirror); + fdb = esw_get_prio_table(esw, attr->chain, attr->prio, !!split); if (IS_ERR(fdb)) { rule = ERR_CAST(fdb); goto err_esw_get; @@ -182,7 +182,7 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw, return rule; err_add_rule: - esw_put_prio_table(esw, attr->chain, attr->prio, !!mirror); + esw_put_prio_table(esw, attr->chain, attr->prio, !!split); err_esw_get: if (attr->dest_chain) esw_put_prio_table(esw, attr->dest_chain, 1, 0); @@ -216,7 +216,7 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw, } flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST; - for (i = 0; i < attr->mirror_count; i++) { + for (i = 0; i < attr->split_count; i++) { dest[i].type = MLX5_FLOW_DESTINATION_TYPE_VPORT; dest[i].vport.num = attr->out_rep[i]->vport; dest[i].vport.vhca_id = @@ -270,7 +270,7 @@ __mlx5_eswitch_del_rule(struct mlx5_eswitch *esw, struct mlx5_esw_flow_attr *attr, bool fwd_rule) { - bool mirror = (attr->mirror_count > 0); + bool split = (attr->split_count > 0); mlx5_del_flow_rules(rule); esw->offloads.num_flows--; @@ -279,7 +279,7 @@ __mlx5_eswitch_del_rule(struct mlx5_eswitch *esw, esw_put_prio_table(esw, attr->chain, attr->prio, 1); esw_put_prio_table(esw, attr->chain, attr->prio, 0); } else { - esw_put_prio_table(esw, attr->chain, attr->prio, !!mirror); + esw_put_prio_table(esw, attr->chain, attr->prio, !!split); if (attr->dest_chain) esw_put_prio_table(esw, attr->dest_chain, 1, 0); } -- cgit v1.2.3-55-g7522 From df65a573eab28059e5dd1d5c32ad270dcf9cbf24 Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Sat, 1 Dec 2018 09:40:43 +0200 Subject: net/mlx5e: Refactor eswitch flow attr for destination specific properties Currently the eswitch flow attr structure stores each destination specific property in its own specific array. Group them in an array of destination structures as a pre-step towards adding additional destination specific field properties. Signed-off-by: Eli Britstein Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 13 ++++++++----- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 6 ++++-- .../net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 16 ++++++++-------- 3 files changed, 20 insertions(+), 15 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 342f807543a0..b24b758f467d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -939,8 +939,8 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, } out_priv = netdev_priv(encap_dev); rpriv = out_priv->ppriv; - attr->out_rep[attr->out_count] = rpriv->rep; - attr->out_mdev[attr->out_count++] = out_priv->mdev; + attr->dests[attr->out_count].rep = rpriv->rep; + attr->dests[attr->out_count++].mdev = out_priv->mdev; } err = mlx5_eswitch_add_vlan_action(esw, attr); @@ -2468,8 +2468,9 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, MLX5_FLOW_CONTEXT_ACTION_COUNT; out_priv = netdev_priv(out_dev); rpriv = out_priv->ppriv; - attr->out_rep[attr->out_count] = rpriv->rep; - attr->out_mdev[attr->out_count++] = out_priv->mdev; + attr->dests[attr->out_count].rep = rpriv->rep; + attr->dests[attr->out_count].mdev = out_priv->mdev; + attr->out_count++; } else if (encap) { parse_attr->mirred_ifindex = out_dev->ifindex; parse_attr->tun_info = *info; @@ -2477,7 +2478,9 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, action |= MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT | MLX5_FLOW_CONTEXT_ACTION_FWD_DEST | MLX5_FLOW_CONTEXT_ACTION_COUNT; - /* attr->out_rep is resolved when we handle encap */ + /* attr->dests[].rep is resolved when we + * handle encap + */ } else if (parse_attr->filter_dev != priv->netdev) { /* All mlx5 devices are called to configure * high level device filters. Therefore, the diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h index c28ff6487dfd..5468975a1bdb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h @@ -283,8 +283,6 @@ enum mlx5_flow_match_level { struct mlx5_esw_flow_attr { struct mlx5_eswitch_rep *in_rep; - struct mlx5_eswitch_rep *out_rep[MLX5_MAX_FLOW_FWD_VPORTS]; - struct mlx5_core_dev *out_mdev[MLX5_MAX_FLOW_FWD_VPORTS]; struct mlx5_core_dev *in_mdev; int split_count; @@ -297,6 +295,10 @@ struct mlx5_esw_flow_attr { u8 total_vlan; bool vlan_handled; u32 encap_id; + struct { + struct mlx5_eswitch_rep *rep; + struct mlx5_core_dev *mdev; + } dests[MLX5_MAX_FLOW_FWD_VPORTS]; u32 mod_hdr_id; u8 match_level; struct mlx5_fc *counter; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 63ac9d3d1690..7cbe602129b7 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -122,9 +122,9 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw, } else { for (j = attr->split_count; j < attr->out_count; j++) { dest[i].type = MLX5_FLOW_DESTINATION_TYPE_VPORT; - dest[i].vport.num = attr->out_rep[j]->vport; + dest[i].vport.num = attr->dests[j].rep->vport; dest[i].vport.vhca_id = - MLX5_CAP_GEN(attr->out_mdev[j], vhca_id); + MLX5_CAP_GEN(attr->dests[j].mdev, vhca_id); if (MLX5_CAP_ESW(esw->dev, merged_eswitch)) dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_VHCA_ID; @@ -218,9 +218,9 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw, flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST; for (i = 0; i < attr->split_count; i++) { dest[i].type = MLX5_FLOW_DESTINATION_TYPE_VPORT; - dest[i].vport.num = attr->out_rep[i]->vport; + dest[i].vport.num = attr->dests[i].rep->vport; dest[i].vport.vhca_id = - MLX5_CAP_GEN(attr->out_mdev[i], vhca_id); + MLX5_CAP_GEN(attr->dests[i].mdev, vhca_id); if (MLX5_CAP_ESW(esw->dev, merged_eswitch)) dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_VHCA_ID; } @@ -327,7 +327,7 @@ esw_vlan_action_get_vport(struct mlx5_esw_flow_attr *attr, bool push, bool pop) struct mlx5_eswitch_rep *in_rep, *out_rep, *vport = NULL; in_rep = attr->in_rep; - out_rep = attr->out_rep[0]; + out_rep = attr->dests[0].rep; if (push) vport = in_rep; @@ -348,7 +348,7 @@ static int esw_add_vlan_action_check(struct mlx5_esw_flow_attr *attr, goto out_notsupp; in_rep = attr->in_rep; - out_rep = attr->out_rep[0]; + out_rep = attr->dests[0].rep; if (push && in_rep->vport == FDB_UPLINK_VPORT) goto out_notsupp; @@ -400,7 +400,7 @@ int mlx5_eswitch_add_vlan_action(struct mlx5_eswitch *esw, if (!push && !pop && fwd) { /* tracks VF --> wire rules without vlan push action */ - if (attr->out_rep[0]->vport == FDB_UPLINK_VPORT) { + if (attr->dests[0].rep->vport == FDB_UPLINK_VPORT) { vport->vlan_refcount++; attr->vlan_handled = true; } @@ -460,7 +460,7 @@ int mlx5_eswitch_del_vlan_action(struct mlx5_eswitch *esw, if (!push && !pop && fwd) { /* tracks VF --> wire rules without vlan push action */ - if (attr->out_rep[0]->vport == FDB_UPLINK_VPORT) + if (attr->dests[0].rep->vport == FDB_UPLINK_VPORT) vport->vlan_refcount--; return 0; -- cgit v1.2.3-55-g7522 From f493f15534ecb5c2c295ba49a2add5822d05ae19 Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Sat, 1 Dec 2018 09:40:43 +0200 Subject: net/mlx5e: Move flow attr reformat action bit to per dest flags Flow attr reformat action bit is moved from the global action bits to a per destination flags field, as a pre-step for adding additional flags to support encapsulation properties per destination, with no functionality change. Signed-off-by: Eli Britstein Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 37 ++++++++++++---------- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 5 +++ .../ethernet/mellanox/mlx5/core/eswitch_offloads.c | 7 ++-- 3 files changed, 30 insertions(+), 19 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index b24b758f467d..e48fbb7d9735 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -908,6 +908,7 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, struct mlx5e_rep_priv *rpriv; struct mlx5e_priv *out_priv; int err = 0, encap_err = 0; + int out_index; /* if prios are not supported, keep the old behaviour of using same prio * for all offloaded rules. @@ -927,7 +928,10 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, goto err_max_prio_chain; } - if (attr->action & MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT) { + for (out_index = 0; out_index < MLX5_MAX_FLOW_FWD_VPORTS; out_index++) { + if (!(attr->dests[out_index].flags & MLX5_ESW_DEST_ENCAP)) + continue; + out_dev = __dev_get_by_index(dev_net(priv->netdev), attr->parse_attr->mirred_ifindex); encap_err = mlx5e_attach_encap(priv, &parse_attr->tun_info, @@ -991,8 +995,11 @@ err_create_counter: err_mod_hdr: mlx5_eswitch_del_vlan_action(esw, attr); err_add_vlan: - if (attr->action & MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT) - mlx5e_detach_encap(priv, flow); + for (out_index = 0; out_index < MLX5_MAX_FLOW_FWD_VPORTS; out_index++) + if (attr->dests[out_index].flags & MLX5_ESW_DEST_ENCAP) { + mlx5e_detach_encap(priv, flow); + break; + } err_attach_encap: err_max_prio_chain: return err; @@ -1004,6 +1011,7 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv, struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; struct mlx5_esw_flow_attr *attr = flow->esw_attr; struct mlx5_esw_flow_attr slow_attr; + int out_index; if (flow->flags & MLX5E_TC_FLOW_OFFLOADED) { if (flow->flags & MLX5E_TC_FLOW_SLOW) @@ -1014,10 +1022,12 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv, mlx5_eswitch_del_vlan_action(esw, attr); - if (attr->action & MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT) { - mlx5e_detach_encap(priv, flow); - kvfree(attr->parse_attr); - } + for (out_index = 0; out_index < MLX5_MAX_FLOW_FWD_VPORTS; out_index++) + if (attr->dests[out_index].flags & MLX5_ESW_DEST_ENCAP) { + mlx5e_detach_encap(priv, flow); + break; + } + kvfree(attr->parse_attr); if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) mlx5e_detach_mod_hdr(priv, flow); @@ -2461,11 +2471,11 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, return -EOPNOTSUPP; } + action |= MLX5_FLOW_CONTEXT_ACTION_FWD_DEST | + MLX5_FLOW_CONTEXT_ACTION_COUNT; if (switchdev_port_same_parent_id(priv->netdev, out_dev) || is_merged_eswitch_dev(priv, out_dev)) { - action |= MLX5_FLOW_CONTEXT_ACTION_FWD_DEST | - MLX5_FLOW_CONTEXT_ACTION_COUNT; out_priv = netdev_priv(out_dev); rpriv = out_priv->ppriv; attr->dests[attr->out_count].rep = rpriv->rep; @@ -2475,9 +2485,8 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, parse_attr->mirred_ifindex = out_dev->ifindex; parse_attr->tun_info = *info; attr->parse_attr = parse_attr; - action |= MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT | - MLX5_FLOW_CONTEXT_ACTION_FWD_DEST | - MLX5_FLOW_CONTEXT_ACTION_COUNT; + attr->dests[attr->out_count].flags |= + MLX5_ESW_DEST_ENCAP; /* attr->dests[].rep is resolved when we * handle encap */ @@ -2657,10 +2666,6 @@ mlx5e_add_fdb_flow(struct mlx5e_priv *priv, if (err) goto err_free; - if (!(flow->esw_attr->action & - MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT)) - kvfree(parse_attr); - *__flow = flow; return 0; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h index 5468975a1bdb..39363d4662b3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h @@ -281,6 +281,10 @@ enum mlx5_flow_match_level { /* current maximum for flow based vport multicasting */ #define MLX5_MAX_FLOW_FWD_VPORTS 2 +enum { + MLX5_ESW_DEST_ENCAP = BIT(0), +}; + struct mlx5_esw_flow_attr { struct mlx5_eswitch_rep *in_rep; struct mlx5_core_dev *in_mdev; @@ -296,6 +300,7 @@ struct mlx5_esw_flow_attr { bool vlan_handled; u32 encap_id; struct { + u32 flags; struct mlx5_eswitch_rep *rep; struct mlx5_core_dev *mdev; } dests[MLX5_MAX_FLOW_FWD_VPORTS]; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 7cbe602129b7..0387b5068be6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -128,6 +128,10 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw, if (MLX5_CAP_ESW(esw->dev, merged_eswitch)) dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_VHCA_ID; + if (attr->dests[j].flags & MLX5_ESW_DEST_ENCAP) { + flow_act.action |= MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT; + flow_act.reformat_id = attr->encap_id; + } i++; } } @@ -164,9 +168,6 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw, if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) flow_act.modify_id = attr->mod_hdr_id; - if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT) - flow_act.reformat_id = attr->encap_id; - fdb = esw_get_prio_table(esw, attr->chain, attr->prio, !!split); if (IS_ERR(fdb)) { rule = ERR_CAST(fdb); -- cgit v1.2.3-55-g7522 From a18e879d4e45b6aef73f42c765cd0da4b26f4260 Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Tue, 4 Dec 2018 08:06:41 +0200 Subject: net/mlx5e: Annul encap action ordering requirement Currently a FW syndrome is emitted if the driver configures a multi-destination FTE where the first destination is a tunneled uplink port and the second destination is a local vPort. Support this scenario by creating a multi-destination FTE using the firmware's extended destination capabilities. Signed-off-by: Eli Britstein Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 0387b5068be6..bb96c4661e26 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -131,6 +131,8 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw, if (attr->dests[j].flags & MLX5_ESW_DEST_ENCAP) { flow_act.action |= MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT; flow_act.reformat_id = attr->encap_id; + dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_REFORMAT_ID; + dest[i].vport.reformat_id = attr->encap_id; } i++; } -- cgit v1.2.3-55-g7522 From 38c9d2697b9c0d928875ea45b5e2fd706df7ad7e Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Thu, 22 Nov 2018 14:04:33 +0200 Subject: net/mlx5e: Replace the split logic with extended destination Currently the FTE encap flag applies to all destinations. To support mirroring encapsulated traffic to a local port the driver split the two destinations to two flow table entries: Table#0: - FWD to the local vport - Goto table#1 Table#1: - Encap and FWD to wire The firmware extended destination capabilities enable the driver to set an encapsulation flag per destination. Remove the split logic and use the extended destination mechanism instead. Note that split technique is still required for pedit and VLAN push scenarios. Signed-off-by: Eli Britstein Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 1 - 1 file changed, 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index e48fbb7d9735..1e3688a0c4f9 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -2513,7 +2513,6 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, encap = true; else return -EOPNOTSUPP; - attr->split_count = attr->out_count; continue; } -- cgit v1.2.3-55-g7522 From 1cc26d74bb98bd9693f0e630ee1cac87d5ba9f20 Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Mon, 26 Nov 2018 07:24:08 +0200 Subject: net/mlx5e: Support header rewrite actions with remote port mirroring A rule with the following actions is split to a two level FDB: 1. Forward to local mirror vport 2. Header rewrite 3. Forward to local vport In the first level flow table, forward the packet to the local port and forward the packet to the second level flow table for header rewrite and local port forwarding. This configuration fails when mirroring to a remote encapsulated destination because currently an FTE cannot support encap and table destinations. Use the extended destination capabilities to configure the first level flow table with a multi-destination FTE to the uplink and second level table and the second level flow table for the header rewrite and local port forwarding. Signed-off-by: Eli Britstein Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 5 +++-- drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 4 ++++ 2 files changed, 7 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 1e3688a0c4f9..4030462f56dc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -943,8 +943,8 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, } out_priv = netdev_priv(encap_dev); rpriv = out_priv->ppriv; - attr->dests[attr->out_count].rep = rpriv->rep; - attr->dests[attr->out_count++].mdev = out_priv->mdev; + attr->dests[out_index].rep = rpriv->rep; + attr->dests[out_index].mdev = out_priv->mdev; } err = mlx5_eswitch_add_vlan_action(esw, attr); @@ -2487,6 +2487,7 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, attr->parse_attr = parse_attr; attr->dests[attr->out_count].flags |= MLX5_ESW_DEST_ENCAP; + attr->out_count++; /* attr->dests[].rep is resolved when we * handle encap */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index bb96c4661e26..b0652e44a68d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -226,6 +226,10 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw, MLX5_CAP_GEN(attr->dests[i].mdev, vhca_id); if (MLX5_CAP_ESW(esw->dev, merged_eswitch)) dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_VHCA_ID; + if (attr->dests[i].flags & MLX5_ESW_DEST_ENCAP) { + dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_REFORMAT_ID; + dest[i].vport.reformat_id = attr->encap_id; + } } dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE; dest[i].ft = fwd_fdb, -- cgit v1.2.3-55-g7522 From 98b66cb1c9ff6ffbc6133ed649b2731638520992 Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Sun, 18 Nov 2018 13:16:52 +0200 Subject: net/mlx5e: Change parse attr struct to accommodate multiple tunnel infos Currently the driver can support only a single TC tunnel_set action. Change the tunnel info fields to arrays, as a pre-step to support multiple encapsulations for a single flow, with no functional change. Signed-off-by: Eli Britstein Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 4030462f56dc..67b3c6f1aeae 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -94,12 +94,12 @@ struct mlx5e_tc_flow { }; struct mlx5e_tc_flow_parse_attr { - struct ip_tunnel_info tun_info; + struct ip_tunnel_info tun_info[MLX5_MAX_FLOW_FWD_VPORTS]; struct net_device *filter_dev; struct mlx5_flow_spec spec; int num_mod_hdr_actions; void *mod_hdr_actions; - int mirred_ifindex; + int mirred_ifindex[MLX5_MAX_FLOW_FWD_VPORTS]; }; #define MLX5E_TC_TABLE_NUM_GROUPS 4 @@ -571,7 +571,7 @@ static int mlx5e_hairpin_flow_add(struct mlx5e_priv *priv, struct mlx5e_tc_flow_parse_attr *parse_attr, struct netlink_ext_ack *extack) { - int peer_ifindex = parse_attr->mirred_ifindex; + int peer_ifindex = parse_attr->mirred_ifindex[0]; struct mlx5_hairpin_params params; struct mlx5_core_dev *peer_mdev; struct mlx5e_hairpin_entry *hpe; @@ -933,8 +933,8 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, continue; out_dev = __dev_get_by_index(dev_net(priv->netdev), - attr->parse_attr->mirred_ifindex); - encap_err = mlx5e_attach_encap(priv, &parse_attr->tun_info, + attr->parse_attr->mirred_ifindex[0]); + encap_err = mlx5e_attach_encap(priv, &parse_attr->tun_info[0], out_dev, &encap_dev, flow, extack); if (encap_err && encap_err != -EAGAIN) { @@ -2220,7 +2220,7 @@ static int parse_tc_nic_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, if (priv->netdev->netdev_ops == peer_dev->netdev_ops && same_hw_devs(priv, netdev_priv(peer_dev))) { - parse_attr->mirred_ifindex = peer_dev->ifindex; + parse_attr->mirred_ifindex[0] = peer_dev->ifindex; flow->flags |= MLX5E_TC_FLOW_HAIRPIN; action |= MLX5_FLOW_CONTEXT_ACTION_FWD_DEST | MLX5_FLOW_CONTEXT_ACTION_COUNT; @@ -2482,8 +2482,8 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, attr->dests[attr->out_count].mdev = out_priv->mdev; attr->out_count++; } else if (encap) { - parse_attr->mirred_ifindex = out_dev->ifindex; - parse_attr->tun_info = *info; + parse_attr->mirred_ifindex[0] = out_dev->ifindex; + parse_attr->tun_info[0] = *info; attr->parse_attr = parse_attr; attr->dests[attr->out_count].flags |= MLX5_ESW_DEST_ENCAP; -- cgit v1.2.3-55-g7522 From 79baaec71989afa805a0f2dd532451a6103e11b4 Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Sun, 2 Dec 2018 09:44:13 +0200 Subject: net/mlx5e: Allow association of a flow to multiple encaps Currently a flow can be associated with a single encap entry. The extended destination feature enables the driver to configure multiple encap entries per flow. Change the encap flow association field to array as a pre-step towards supporting multiple encap destinations. Use only the first array element, with no functional change. Signed-off-by: Eli Britstein Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 48 +++++++++++++++++++++---- 1 file changed, 41 insertions(+), 7 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 67b3c6f1aeae..58b83489a6d6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -78,13 +78,39 @@ enum { #define MLX5E_TC_MAX_SPLITS 1 +/* Helper struct for accessing a struct containing list_head array. + * Containing struct + * |- Helper array + * [0] Helper item 0 + * |- list_head item 0 + * |- index (0) + * [1] Helper item 1 + * |- list_head item 1 + * |- index (1) + * To access the containing struct from one of the list_head items: + * 1. Get the helper item from the list_head item using + * helper item = + * container_of(list_head item, helper struct type, list_head field) + * 2. Get the contining struct from the helper item and its index in the array: + * containing struct = + * container_of(helper item, containing struct type, helper field[index]) + */ +struct encap_flow_item { + struct list_head list; + int index; +}; + struct mlx5e_tc_flow { struct rhash_head node; struct mlx5e_priv *priv; u64 cookie; u16 flags; struct mlx5_flow_handle *rule[MLX5E_TC_MAX_SPLITS + 1]; - struct list_head encap; /* flows sharing the same encap ID */ + /* Flow can be associated with multiple encap IDs. + * The number of encaps is bounded by the number of supported + * destinations. + */ + struct encap_flow_item encaps[MLX5_MAX_FLOW_FWD_VPORTS]; struct list_head mod_hdr; /* flows sharing the same mod hdr ID */ struct list_head hairpin; /* flows sharing the same hairpin */ union { @@ -1043,6 +1069,7 @@ void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv, struct mlx5_esw_flow_attr slow_attr, *esw_attr; struct mlx5_flow_handle *rule; struct mlx5_flow_spec *spec; + struct encap_flow_item *efi; struct mlx5e_tc_flow *flow; int err; @@ -1059,7 +1086,8 @@ void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv, e->flags |= MLX5_ENCAP_ENTRY_VALID; mlx5e_rep_queue_neigh_stats_work(priv); - list_for_each_entry(flow, &e->flows, encap) { + list_for_each_entry(efi, &e->flows, list) { + flow = container_of(efi, struct mlx5e_tc_flow, encaps[efi->index]); esw_attr = flow->esw_attr; esw_attr->encap_id = e->encap_id; spec = &esw_attr->parse_attr->spec; @@ -1086,10 +1114,12 @@ void mlx5e_tc_encap_flows_del(struct mlx5e_priv *priv, struct mlx5_esw_flow_attr slow_attr; struct mlx5_flow_handle *rule; struct mlx5_flow_spec *spec; + struct encap_flow_item *efi; struct mlx5e_tc_flow *flow; int err; - list_for_each_entry(flow, &e->flows, encap) { + list_for_each_entry(efi, &e->flows, list) { + flow = container_of(efi, struct mlx5e_tc_flow, encaps[efi->index]); spec = &flow->esw_attr->parse_attr->spec; /* update from encap rule to slow path rule */ @@ -1142,9 +1172,12 @@ void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe) return; list_for_each_entry(e, &nhe->encap_list, encap_list) { + struct encap_flow_item *efi; if (!(e->flags & MLX5_ENCAP_ENTRY_VALID)) continue; - list_for_each_entry(flow, &e->flows, encap) { + list_for_each_entry(efi, &e->flows, list) { + flow = container_of(efi, struct mlx5e_tc_flow, + encaps[efi->index]); if (flow->flags & MLX5E_TC_FLOW_OFFLOADED) { counter = mlx5e_tc_get_counter(flow); mlx5_fc_query_cached(counter, &bytes, &packets, &lastuse); @@ -1176,9 +1209,9 @@ void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe) static void mlx5e_detach_encap(struct mlx5e_priv *priv, struct mlx5e_tc_flow *flow) { - struct list_head *next = flow->encap.next; + struct list_head *next = flow->encaps[0].list.next; - list_del(&flow->encap); + list_del(&flow->encaps[0].list); if (list_empty(next)) { struct mlx5e_encap_entry *e; @@ -2338,7 +2371,8 @@ static int mlx5e_attach_encap(struct mlx5e_priv *priv, hash_add_rcu(esw->offloads.encap_tbl, &e->encap_hlist, hash_key); attach_flow: - list_add(&flow->encap, &e->flows); + list_add(&flow->encaps[0].list, &e->flows); + flow->encaps[0].index = 0; *encap_dev = e->out_dev; if (e->flags & MLX5_ENCAP_ENTRY_VALID) attr->encap_id = e->encap_id; -- cgit v1.2.3-55-g7522 From 8c4dc42bf6e4ffeda49cf5e26bfc991b548fc0aa Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Sun, 18 Nov 2018 14:21:46 +0200 Subject: net/mlx5e: Support multiple encapsulations for a TC flow Currently a flow is associated with a single encap structure. The FW extended destination features enables the driver to associate a flow with multiple encap instances. Change the encap id field from a flow scope to a per destination value in the flow attributes struct. Use the encaps array to associate a flow table entry with multiple encap entries. Update the neigh logic to offload only if all encapsulations used in a flow are connected, and un-offload upon the first one disconnected. Note that the driver can now support up to two encap destinations. Signed-off-by: Eli Britstein Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 85 ++++++++++++++-------- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 3 +- .../ethernet/mellanox/mlx5/core/eswitch_offloads.c | 7 +- 3 files changed, 62 insertions(+), 33 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 58b83489a6d6..0921213561cb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -843,14 +843,15 @@ static void mlx5e_tc_del_nic_flow(struct mlx5e_priv *priv, } static void mlx5e_detach_encap(struct mlx5e_priv *priv, - struct mlx5e_tc_flow *flow); + struct mlx5e_tc_flow *flow, int out_index); static int mlx5e_attach_encap(struct mlx5e_priv *priv, struct ip_tunnel_info *tun_info, struct net_device *mirred_dev, struct net_device **encap_dev, struct mlx5e_tc_flow *flow, - struct netlink_ext_ack *extack); + struct netlink_ext_ack *extack, + int out_index); static struct mlx5_flow_handle * mlx5e_tc_offload_fdb_rules(struct mlx5_eswitch *esw, @@ -955,18 +956,22 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, } for (out_index = 0; out_index < MLX5_MAX_FLOW_FWD_VPORTS; out_index++) { + int mirred_ifindex; + if (!(attr->dests[out_index].flags & MLX5_ESW_DEST_ENCAP)) continue; + mirred_ifindex = attr->parse_attr->mirred_ifindex[out_index]; out_dev = __dev_get_by_index(dev_net(priv->netdev), - attr->parse_attr->mirred_ifindex[0]); - encap_err = mlx5e_attach_encap(priv, &parse_attr->tun_info[0], - out_dev, &encap_dev, flow, - extack); - if (encap_err && encap_err != -EAGAIN) { - err = encap_err; + mirred_ifindex); + err = mlx5e_attach_encap(priv, + &parse_attr->tun_info[out_index], + out_dev, &encap_dev, flow, + extack, out_index); + if (err && err != -EAGAIN) goto err_attach_encap; - } + if (err == -EAGAIN) + encap_err = err; out_priv = netdev_priv(encap_dev); rpriv = out_priv->ppriv; attr->dests[out_index].rep = rpriv->rep; @@ -1022,10 +1027,8 @@ err_mod_hdr: mlx5_eswitch_del_vlan_action(esw, attr); err_add_vlan: for (out_index = 0; out_index < MLX5_MAX_FLOW_FWD_VPORTS; out_index++) - if (attr->dests[out_index].flags & MLX5_ESW_DEST_ENCAP) { - mlx5e_detach_encap(priv, flow); - break; - } + if (attr->dests[out_index].flags & MLX5_ESW_DEST_ENCAP) + mlx5e_detach_encap(priv, flow, out_index); err_attach_encap: err_max_prio_chain: return err; @@ -1049,10 +1052,8 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv, mlx5_eswitch_del_vlan_action(esw, attr); for (out_index = 0; out_index < MLX5_MAX_FLOW_FWD_VPORTS; out_index++) - if (attr->dests[out_index].flags & MLX5_ESW_DEST_ENCAP) { - mlx5e_detach_encap(priv, flow); - break; - } + if (attr->dests[out_index].flags & MLX5_ESW_DEST_ENCAP) + mlx5e_detach_encap(priv, flow, out_index); kvfree(attr->parse_attr); if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) @@ -1087,11 +1088,30 @@ void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv, mlx5e_rep_queue_neigh_stats_work(priv); list_for_each_entry(efi, &e->flows, list) { + bool all_flow_encaps_valid = true; + int i; + flow = container_of(efi, struct mlx5e_tc_flow, encaps[efi->index]); esw_attr = flow->esw_attr; - esw_attr->encap_id = e->encap_id; spec = &esw_attr->parse_attr->spec; + esw_attr->dests[efi->index].encap_id = e->encap_id; + esw_attr->dests[efi->index].flags |= MLX5_ESW_DEST_ENCAP_VALID; + /* Flow can be associated with multiple encap entries. + * Before offloading the flow verify that all of them have + * a valid neighbour. + */ + for (i = 0; i < MLX5_MAX_FLOW_FWD_VPORTS; i++) { + if (!(esw_attr->dests[i].flags & MLX5_ESW_DEST_ENCAP)) + continue; + if (!(esw_attr->dests[i].flags & MLX5_ESW_DEST_ENCAP_VALID)) { + all_flow_encaps_valid = false; + break; + } + } + /* Do not offload flows with unresolved neighbors */ + if (!all_flow_encaps_valid) + continue; /* update from slow path rule to encap rule */ rule = mlx5e_tc_offload_fdb_rules(esw, flow, spec, esw_attr); if (IS_ERR(rule)) { @@ -1124,6 +1144,8 @@ void mlx5e_tc_encap_flows_del(struct mlx5e_priv *priv, /* update from encap rule to slow path rule */ rule = mlx5e_tc_offload_to_slow_path(esw, flow, spec, &slow_attr); + /* mark the flow's encap dest as non-valid */ + flow->esw_attr->dests[efi->index].flags &= ~MLX5_ESW_DEST_ENCAP_VALID; if (IS_ERR(rule)) { err = PTR_ERR(rule); @@ -1207,11 +1229,11 @@ void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe) } static void mlx5e_detach_encap(struct mlx5e_priv *priv, - struct mlx5e_tc_flow *flow) + struct mlx5e_tc_flow *flow, int out_index) { - struct list_head *next = flow->encaps[0].list.next; + struct list_head *next = flow->encaps[out_index].list.next; - list_del(&flow->encaps[0].list); + list_del(&flow->encaps[out_index].list); if (list_empty(next)) { struct mlx5e_encap_entry *e; @@ -2324,7 +2346,8 @@ static int mlx5e_attach_encap(struct mlx5e_priv *priv, struct net_device *mirred_dev, struct net_device **encap_dev, struct mlx5e_tc_flow *flow, - struct netlink_ext_ack *extack) + struct netlink_ext_ack *extack, + int out_index) { struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; unsigned short family = ip_tunnel_info_af(tun_info); @@ -2371,13 +2394,15 @@ static int mlx5e_attach_encap(struct mlx5e_priv *priv, hash_add_rcu(esw->offloads.encap_tbl, &e->encap_hlist, hash_key); attach_flow: - list_add(&flow->encaps[0].list, &e->flows); - flow->encaps[0].index = 0; + list_add(&flow->encaps[out_index].list, &e->flows); + flow->encaps[out_index].index = out_index; *encap_dev = e->out_dev; - if (e->flags & MLX5_ENCAP_ENTRY_VALID) - attr->encap_id = e->encap_id; - else + if (e->flags & MLX5_ENCAP_ENTRY_VALID) { + attr->dests[out_index].encap_id = e->encap_id; + attr->dests[out_index].flags |= MLX5_ESW_DEST_ENCAP_VALID; + } else { err = -EAGAIN; + } return err; @@ -2516,8 +2541,10 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, attr->dests[attr->out_count].mdev = out_priv->mdev; attr->out_count++; } else if (encap) { - parse_attr->mirred_ifindex[0] = out_dev->ifindex; - parse_attr->tun_info[0] = *info; + parse_attr->mirred_ifindex[attr->out_count] = + out_dev->ifindex; + parse_attr->tun_info[attr->out_count] = *info; + encap = false; attr->parse_attr = parse_attr; attr->dests[attr->out_count].flags |= MLX5_ESW_DEST_ENCAP; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h index 39363d4662b3..87c9dea9bccf 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h @@ -283,6 +283,7 @@ enum mlx5_flow_match_level { enum { MLX5_ESW_DEST_ENCAP = BIT(0), + MLX5_ESW_DEST_ENCAP_VALID = BIT(1), }; struct mlx5_esw_flow_attr { @@ -298,11 +299,11 @@ struct mlx5_esw_flow_attr { u8 vlan_prio[MLX5_FS_VLAN_DEPTH]; u8 total_vlan; bool vlan_handled; - u32 encap_id; struct { u32 flags; struct mlx5_eswitch_rep *rep; struct mlx5_core_dev *mdev; + u32 encap_id; } dests[MLX5_MAX_FLOW_FWD_VPORTS]; u32 mod_hdr_id; u8 match_level; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index b0652e44a68d..bde1fb8c284b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -130,9 +130,10 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw, MLX5_FLOW_DEST_VPORT_VHCA_ID; if (attr->dests[j].flags & MLX5_ESW_DEST_ENCAP) { flow_act.action |= MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT; - flow_act.reformat_id = attr->encap_id; + flow_act.reformat_id = attr->dests[j].encap_id; dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_REFORMAT_ID; - dest[i].vport.reformat_id = attr->encap_id; + dest[i].vport.reformat_id = + attr->dests[j].encap_id; } i++; } @@ -228,7 +229,7 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw, dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_VHCA_ID; if (attr->dests[i].flags & MLX5_ESW_DEST_ENCAP) { dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_REFORMAT_ID; - dest[i].vport.reformat_id = attr->encap_id; + dest[i].vport.reformat_id = attr->dests[i].encap_id; } } dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE; -- cgit v1.2.3-55-g7522 From 939de57d30344ce728b0de61be87984e75af420e Mon Sep 17 00:00:00 2001 From: Daniel Jurgens Date: Mon, 5 Nov 2018 16:05:37 -0600 Subject: net/mlx5e: Use CQE padding for Ethernet CQs Writing 64B CQEs to 128B cache lines results in a RMW operation. Padding the CQEs to 128B if possible improves performance on 128B cache line systems like PPC. Testing on PPC showed up to a 24% improvement in small packet throughput vs the default behavior, depending on the workload and system topology. Signed-off-by: Daniel Jurgens Reviewed-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 2 ++ drivers/net/ethernet/mellanox/mlx5/core/wq.c | 3 ++- drivers/net/ethernet/mellanox/mlx5/core/wq.h | 7 ++++++- include/linux/mlx5/cq.h | 10 +++++----- 4 files changed, 15 insertions(+), 7 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 88116a4750b0..2188e5ba908f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -2224,6 +2224,8 @@ static void mlx5e_build_common_cq_param(struct mlx5e_priv *priv, void *cqc = param->cqc; MLX5_SET(cqc, cqc, uar_page, priv->mdev->priv.uar->index); + if (MLX5_CAP_GEN(priv->mdev, cqe_128_always) && cache_line_size() >= 128) + MLX5_SET(cqc, cqc, cqe_sz, CQE_STRIDE_128_PAD); } static void mlx5e_build_rx_cq_param(struct mlx5e_priv *priv, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/wq.c b/drivers/net/ethernet/mellanox/mlx5/core/wq.c index 2dcbf1ebfd6a..953cc8efba69 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/wq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/wq.c @@ -155,7 +155,8 @@ int mlx5_cqwq_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, void *cqc, struct mlx5_cqwq *wq, struct mlx5_wq_ctrl *wq_ctrl) { - u8 log_wq_stride = MLX5_GET(cqc, cqc, cqe_sz) + 6; + /* CQE_STRIDE_128 and CQE_STRIDE_128_PAD both mean 128B stride */ + u8 log_wq_stride = MLX5_GET(cqc, cqc, cqe_sz) == CQE_STRIDE_64 ? 6 : 7; u8 log_wq_sz = MLX5_GET(cqc, cqc, log_cq_size); int err; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/wq.h b/drivers/net/ethernet/mellanox/mlx5/core/wq.h index 9bc2184a46bc..ea934a48c90a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/wq.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/wq.h @@ -179,7 +179,12 @@ static inline u32 mlx5_cqwq_get_ci(struct mlx5_cqwq *wq) static inline struct mlx5_cqe64 *mlx5_cqwq_get_wqe(struct mlx5_cqwq *wq, u32 ix) { - return mlx5_frag_buf_get_wqe(&wq->fbc, ix); + struct mlx5_cqe64 *cqe = mlx5_frag_buf_get_wqe(&wq->fbc, ix); + + /* For 128B CQEs the data is in the last 64B */ + cqe += wq->fbc.log_stride == 7; + + return cqe; } static inline u32 mlx5_cqwq_get_ctr_wrap_cnt(struct mlx5_cqwq *wq, u32 ctr) diff --git a/include/linux/mlx5/cq.h b/include/linux/mlx5/cq.h index 28b757a64029..612c8c2f2466 100644 --- a/include/linux/mlx5/cq.h +++ b/include/linux/mlx5/cq.h @@ -125,9 +125,9 @@ struct mlx5_cq_modify_params { }; enum { - CQE_SIZE_64 = 0, - CQE_SIZE_128 = 1, - CQE_SIZE_128_PAD = 2, + CQE_STRIDE_64 = 0, + CQE_STRIDE_128 = 1, + CQE_STRIDE_128_PAD = 2, }; #define MLX5_MAX_CQ_PERIOD (BIT(__mlx5_bit_sz(cqc, cq_period)) - 1) @@ -135,8 +135,8 @@ enum { static inline int cqe_sz_to_mlx_sz(u8 size, int padding_128_en) { - return padding_128_en ? CQE_SIZE_128_PAD : - size == 64 ? CQE_SIZE_64 : CQE_SIZE_128; + return padding_128_en ? CQE_STRIDE_128_PAD : + size == 64 ? CQE_STRIDE_64 : CQE_STRIDE_128; } static inline void mlx5_cq_set_ci(struct mlx5_core_cq *cq) -- cgit v1.2.3-55-g7522 From 75370eb0d3b802f54600f2fc3ae5255fe9270112 Mon Sep 17 00:00:00 2001 From: Eyal Davidovich Date: Sun, 7 Oct 2018 15:18:37 +0300 Subject: net/mlx5e: Avoid query PPCNT register if not supported by the device PPCNT is not supported if PCAM access reg is supported and ppcnt bit is clear. Signed-off-by: Eyal Davidovich Reviewed-by: Eran Ben Elisha Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_stats.c | 12 ++++++++++++ include/linux/mlx5/mlx5_ifc.h | 4 +++- 2 files changed, 15 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c index 75d30fa637d6..8224f1e062a8 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c @@ -483,6 +483,9 @@ static int mlx5e_grp_802_3_fill_stats(struct mlx5e_priv *priv, u64 *data, return idx; } +#define MLX5_BASIC_PPCNT_SUPPORTED(mdev) \ + (MLX5_CAP_GEN(mdev, pcam_reg) ? MLX5_CAP_PCAM_REG(mdev, ppcnt) : 1) + static void mlx5e_grp_802_3_update_stats(struct mlx5e_priv *priv) { struct mlx5e_pport_stats *pstats = &priv->stats.pport; @@ -491,6 +494,9 @@ static void mlx5e_grp_802_3_update_stats(struct mlx5e_priv *priv) int sz = MLX5_ST_SZ_BYTES(ppcnt_reg); void *out; + if (!MLX5_BASIC_PPCNT_SUPPORTED(mdev)) + return; + MLX5_SET(ppcnt_reg, in, local_port, 1); out = pstats->IEEE_802_3_counters; MLX5_SET(ppcnt_reg, in, grp, MLX5_IEEE_802_3_COUNTERS_GROUP); @@ -603,6 +609,9 @@ static void mlx5e_grp_2819_update_stats(struct mlx5e_priv *priv) int sz = MLX5_ST_SZ_BYTES(ppcnt_reg); void *out; + if (!MLX5_BASIC_PPCNT_SUPPORTED(mdev)) + return; + MLX5_SET(ppcnt_reg, in, local_port, 1); out = pstats->RFC_2819_counters; MLX5_SET(ppcnt_reg, in, grp, MLX5_RFC_2819_COUNTERS_GROUP); @@ -1078,6 +1087,9 @@ static void mlx5e_grp_per_prio_update_stats(struct mlx5e_priv *priv) int prio; void *out; + if (!MLX5_BASIC_PPCNT_SUPPORTED(mdev)) + return; + MLX5_SET(ppcnt_reg, in, local_port, 1); MLX5_SET(ppcnt_reg, in, grp, MLX5_PER_PRIORITY_COUNTERS_GROUP); for (prio = 0; prio < NUM_PPORT_PRIO; prio++) { diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index 60c1d49eb40c..c12b0dec2889 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -8283,7 +8283,9 @@ struct mlx5_ifc_pcam_regs_5000_to_507f_bits { u8 port_access_reg_cap_mask_31_to_13[0x13]; u8 pbmc[0x1]; u8 pptb[0x1]; - u8 port_access_reg_cap_mask_10_to_0[0xb]; + u8 port_access_reg_cap_mask_10_to_09[0x2]; + u8 ppcnt[0x1]; + u8 port_access_reg_cap_mask_07_to_00[0x8]; }; struct mlx5_ifc_pcam_reg_bits { -- cgit v1.2.3-55-g7522 From 2f8bc4917a86f4504f85f0e24d67d17bd2a9bfc1 Mon Sep 17 00:00:00 2001 From: Eyal Davidovich Date: Mon, 8 Oct 2018 12:16:01 +0300 Subject: net/mlx5e: Monitor counters commands support new file monitor_stats.c for the new API. add arm_monitor_counter new command support. add set_monitor_counter new command support. The device can monitor specific counters and provide an event to notify when these counters are changed. The monitoring is done in best effort manner where the minimum notification period is 200 ms, however when the device is loaded, the notification might be delayed. To configure the required counters to be monitored, the SET_MONITOR_COUNTER command shall be used with a list of counters to be monitored. The device firmware can monitor up to HCA_CAP.max_num_of_monitor_counters. The configuration is done based on counter type (such as ppcnt, q counter, etc) and additional param according to the type of counter selected. Upon monitor counter change, the device will generate Monitor_Counter_Change event. The device will not generate new events unless the driver re-arms the monitoring functionality, using the ARM_MONITOR_COUNTER command. Signed-off-by: Eyal Davidovich Reviewed-by: Saeed Mahameed Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/Makefile | 2 +- .../ethernet/mellanox/mlx5/core/en/monitor_stats.c | 126 +++++++++++++++++++++ .../ethernet/mellanox/mlx5/core/en/monitor_stats.h | 12 ++ 3 files changed, 139 insertions(+), 1 deletion(-) create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en/monitor_stats.c create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en/monitor_stats.h (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile index 40764d87413a..9678051b8ff1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile +++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile @@ -22,7 +22,7 @@ mlx5_core-y := main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \ # mlx5_core-$(CONFIG_MLX5_CORE_EN) += en_main.o en_common.o en_fs.o en_ethtool.o \ en_tx.o en_rx.o en_dim.o en_txrx.o en/xdp.o en_stats.o \ - en_selftest.o en/port.o + en_selftest.o en/port.o en/monitor_stats.o # # Netdev extra diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/monitor_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en/monitor_stats.c new file mode 100644 index 000000000000..0f010efa1375 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/monitor_stats.c @@ -0,0 +1,126 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* Copyright (c) 2018 Mellanox Technologies. */ + +#include "en.h" +#include "monitor_stats.h" + +/* Driver will set the following watch counters list: + * Ppcnt.802_3: + * a_in_range_length_errors Type: 0x0, Counter: 0x0, group_id = N/A + * a_out_of_range_length_field Type: 0x0, Counter: 0x1, group_id = N/A + * a_frame_too_long_errors Type: 0x0, Counter: 0x2, group_id = N/A + * a_frame_check_sequence_errors Type: 0x0, Counter: 0x3, group_id = N/A + * a_alignment_errors Type: 0x0, Counter: 0x4, group_id = N/A + * if_out_discards Type: 0x0, Counter: 0x5, group_id = N/A + * Q_Counters: + * Q[index].rx_out_of_buffer Type: 0x1, Counter: 0x4, group_id = counter_ix + */ + +#define NUM_REQ_PPCNT_COUNTER_S1 MLX5_CMD_SET_MONITOR_NUM_PPCNT_COUNTER_SET1 +#define NUM_REQ_Q_COUNTERS_S1 MLX5_CMD_SET_MONITOR_NUM_Q_COUNTERS_SET1 + +int mlx5e_monitor_counter_supported(struct mlx5e_priv *priv) +{ + struct mlx5_core_dev *mdev = priv->mdev; + + if (!MLX5_CAP_GEN(mdev, max_num_of_monitor_counters)) + return false; + if (MLX5_CAP_PCAM_REG(mdev, ppcnt) && + MLX5_CAP_GEN(mdev, num_ppcnt_monitor_counters) < + NUM_REQ_PPCNT_COUNTER_S1) + return false; + if (MLX5_CAP_GEN(mdev, num_q_monitor_counters) < + NUM_REQ_Q_COUNTERS_S1) + return false; + return true; +} + +void mlx5e_monitor_counter_arm(struct mlx5e_priv *priv) +{ + u32 in[MLX5_ST_SZ_DW(arm_monitor_counter_in)] = {}; + u32 out[MLX5_ST_SZ_DW(arm_monitor_counter_out)] = {}; + + MLX5_SET(arm_monitor_counter_in, in, opcode, + MLX5_CMD_OP_ARM_MONITOR_COUNTER); + mlx5_cmd_exec(priv->mdev, in, sizeof(in), out, sizeof(out)); +} + +static int fill_monitor_counter_ppcnt_set1(int cnt, u32 *in) +{ + enum mlx5_monitor_counter_ppcnt ppcnt_cnt; + + for (ppcnt_cnt = 0; + ppcnt_cnt < NUM_REQ_PPCNT_COUNTER_S1; + ppcnt_cnt++, cnt++) { + MLX5_SET(set_monitor_counter_in, in, + monitor_counter[cnt].type, + MLX5_QUERY_MONITOR_CNT_TYPE_PPCNT); + MLX5_SET(set_monitor_counter_in, in, + monitor_counter[cnt].counter, + ppcnt_cnt); + } + return ppcnt_cnt; +} + +static int fill_monitor_counter_q_counter_set1(int cnt, int q_counter, u32 *in) +{ + MLX5_SET(set_monitor_counter_in, in, + monitor_counter[cnt].type, + MLX5_QUERY_MONITOR_CNT_TYPE_Q_COUNTER); + MLX5_SET(set_monitor_counter_in, in, + monitor_counter[cnt].counter, + MLX5_QUERY_MONITOR_Q_COUNTER_RX_OUT_OF_BUFFER); + MLX5_SET(set_monitor_counter_in, in, + monitor_counter[cnt].counter_group_id, + q_counter); + return 1; +} + +/* check if mlx5e_monitor_counter_supported before calling this function*/ +static void mlx5e_set_monitor_counter(struct mlx5e_priv *priv) +{ + struct mlx5_core_dev *mdev = priv->mdev; + int max_num_of_counters = MLX5_CAP_GEN(mdev, max_num_of_monitor_counters); + int num_q_counters = MLX5_CAP_GEN(mdev, num_q_monitor_counters); + int num_ppcnt_counters = !MLX5_CAP_PCAM_REG(mdev, ppcnt) ? 0 : + MLX5_CAP_GEN(mdev, num_ppcnt_monitor_counters); + u32 in[MLX5_ST_SZ_DW(set_monitor_counter_in)] = {}; + u32 out[MLX5_ST_SZ_DW(set_monitor_counter_out)] = {}; + int q_counter = priv->q_counter; + int cnt = 0; + + if (num_ppcnt_counters >= NUM_REQ_PPCNT_COUNTER_S1 && + max_num_of_counters >= (NUM_REQ_PPCNT_COUNTER_S1 + cnt)) + cnt += fill_monitor_counter_ppcnt_set1(cnt, in); + + if (num_q_counters >= NUM_REQ_Q_COUNTERS_S1 && + max_num_of_counters >= (NUM_REQ_Q_COUNTERS_S1 + cnt) && + q_counter) + cnt += fill_monitor_counter_q_counter_set1(cnt, q_counter, in); + + MLX5_SET(set_monitor_counter_in, in, num_of_counters, cnt); + MLX5_SET(set_monitor_counter_in, in, opcode, + MLX5_CMD_OP_SET_MONITOR_COUNTER); + + mlx5_cmd_exec(mdev, in, sizeof(in), out, sizeof(out)); +} + +/* check if mlx5e_monitor_counter_supported before calling this function*/ +void mlx5e_monitor_counter_init(struct mlx5e_priv *priv) +{ + mlx5e_set_monitor_counter(priv); + mlx5e_monitor_counter_arm(priv); +} + +/* check if mlx5e_monitor_counter_supported before calling this function*/ +void mlx5e_monitor_counter_cleanup(struct mlx5e_priv *priv) +{ + u32 in[MLX5_ST_SZ_DW(set_monitor_counter_in)] = {}; + u32 out[MLX5_ST_SZ_DW(set_monitor_counter_out)] = {}; + + MLX5_SET(set_monitor_counter_in, in, num_of_counters, 0); + MLX5_SET(set_monitor_counter_in, in, opcode, + MLX5_CMD_OP_SET_MONITOR_COUNTER); + + mlx5_cmd_exec(priv->mdev, in, sizeof(in), out, sizeof(out)); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/monitor_stats.h b/drivers/net/ethernet/mellanox/mlx5/core/en/monitor_stats.h new file mode 100644 index 000000000000..e1ac4b3d22fb --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/monitor_stats.h @@ -0,0 +1,12 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* Copyright (c) 2018 Mellanox Technologies. */ + +#ifndef __MLX5_MONITOR_H__ +#define __MLX5_MONITOR_H__ + +int mlx5e_monitor_counter_supported(struct mlx5e_priv *priv); +void mlx5e_monitor_counter_init(struct mlx5e_priv *priv); +void mlx5e_monitor_counter_cleanup(struct mlx5e_priv *priv); +void mlx5e_monitor_counter_arm(struct mlx5e_priv *priv); + +#endif /* __MLX5_MONITOR_H__ */ -- cgit v1.2.3-55-g7522 From 5c7e8bbb025726f9cb9b2bc050509a94262426f8 Mon Sep 17 00:00:00 2001 From: Eyal Davidovich Date: Sat, 20 Oct 2018 16:18:00 +0300 Subject: net/mlx5e: Use monitor counters for update stats - Adding new notifier block (struct mlx5_nb) monitor_counters_nb for handeling MONITOR_COUNTER new event type. - Adding work queue element: monitor_counters_work for re-arm and update stats. - We re-queue the update stat work, only when working over firmware that doesn't support the monitored counters. Signed-off-by: Eyal Davidovich Reviewed-by: Saeed Mahameed Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 3 ++ .../ethernet/mellanox/mlx5/core/en/monitor_stats.c | 47 +++++++++++++++++++++- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 14 +++++-- 3 files changed, 59 insertions(+), 5 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 9ea0f853c3bb..1c4e033b7509 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -685,6 +685,8 @@ struct mlx5e_priv { struct work_struct set_rx_mode_work; struct work_struct tx_timeout_work; struct work_struct update_stats_work; + struct work_struct monitor_counters_work; + struct mlx5_nb monitor_counters_nb; struct mlx5_core_dev *mdev; struct net_device *netdev; @@ -940,6 +942,7 @@ int mlx5e_create_tises(struct mlx5e_priv *priv); void mlx5e_cleanup_nic_tx(struct mlx5e_priv *priv); int mlx5e_close(struct net_device *netdev); int mlx5e_open(struct net_device *netdev); +void mlx5e_update_ndo_stats(struct mlx5e_priv *priv); void mlx5e_queue_update_stats(struct mlx5e_priv *priv); int mlx5e_bits_invert(unsigned long a, int size); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/monitor_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en/monitor_stats.c index 0f010efa1375..2ce420851e77 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/monitor_stats.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/monitor_stats.c @@ -3,6 +3,7 @@ #include "en.h" #include "monitor_stats.h" +#include "lib/eq.h" /* Driver will set the following watch counters list: * Ppcnt.802_3: @@ -45,6 +46,39 @@ void mlx5e_monitor_counter_arm(struct mlx5e_priv *priv) mlx5_cmd_exec(priv->mdev, in, sizeof(in), out, sizeof(out)); } +static void mlx5e_monitor_counters_work(struct work_struct *work) +{ + struct mlx5e_priv *priv = container_of(work, struct mlx5e_priv, + monitor_counters_work); + + mutex_lock(&priv->state_lock); + mlx5e_update_ndo_stats(priv); + mutex_unlock(&priv->state_lock); + mlx5e_monitor_counter_arm(priv); +} + +static int mlx5e_monitor_event_handler(struct notifier_block *nb, + unsigned long event, void *eqe) +{ + struct mlx5e_priv *priv = mlx5_nb_cof(nb, struct mlx5e_priv, + monitor_counters_nb); + queue_work(priv->wq, &priv->monitor_counters_work); + return NOTIFY_OK; +} + +void mlx5e_monitor_counter_start(struct mlx5e_priv *priv) +{ + MLX5_NB_INIT(&priv->monitor_counters_nb, mlx5e_monitor_event_handler, + MONITOR_COUNTER); + mlx5_eq_notifier_register(priv->mdev, &priv->monitor_counters_nb); +} + +static void mlx5e_monitor_counter_stop(struct mlx5e_priv *priv) +{ + mlx5_eq_notifier_unregister(priv->mdev, &priv->monitor_counters_nb); + cancel_work_sync(&priv->monitor_counters_work); +} + static int fill_monitor_counter_ppcnt_set1(int cnt, u32 *in) { enum mlx5_monitor_counter_ppcnt ppcnt_cnt; @@ -108,12 +142,14 @@ static void mlx5e_set_monitor_counter(struct mlx5e_priv *priv) /* check if mlx5e_monitor_counter_supported before calling this function*/ void mlx5e_monitor_counter_init(struct mlx5e_priv *priv) { + INIT_WORK(&priv->monitor_counters_work, mlx5e_monitor_counters_work); + mlx5e_monitor_counter_start(priv); mlx5e_set_monitor_counter(priv); mlx5e_monitor_counter_arm(priv); + queue_work(priv->wq, &priv->update_stats_work); } -/* check if mlx5e_monitor_counter_supported before calling this function*/ -void mlx5e_monitor_counter_cleanup(struct mlx5e_priv *priv) +static void mlx5e_monitor_counter_disable(struct mlx5e_priv *priv) { u32 in[MLX5_ST_SZ_DW(set_monitor_counter_in)] = {}; u32 out[MLX5_ST_SZ_DW(set_monitor_counter_out)] = {}; @@ -124,3 +160,10 @@ void mlx5e_monitor_counter_cleanup(struct mlx5e_priv *priv) mlx5_cmd_exec(priv->mdev, in, sizeof(in), out, sizeof(out)); } + +/* check if mlx5e_monitor_counter_supported before calling this function*/ +void mlx5e_monitor_counter_cleanup(struct mlx5e_priv *priv) +{ + mlx5e_monitor_counter_disable(priv); + mlx5e_monitor_counter_stop(priv); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 2188e5ba908f..59745851b5ba 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -50,6 +50,7 @@ #include "en/port.h" #include "en/xdp.h" #include "lib/eq.h" +#include "en/monitor_stats.h" struct mlx5e_rq_param { u32 rqc[MLX5_ST_SZ_DW(rqc)]; @@ -263,7 +264,7 @@ void mlx5e_update_stats(struct mlx5e_priv *priv) mlx5e_stats_grps[i].update_stats(priv); } -static void mlx5e_update_ndo_stats(struct mlx5e_priv *priv) +void mlx5e_update_ndo_stats(struct mlx5e_priv *priv) { int i; @@ -3459,8 +3460,10 @@ mlx5e_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats) struct mlx5e_vport_stats *vstats = &priv->stats.vport; struct mlx5e_pport_stats *pstats = &priv->stats.pport; - /* update HW stats in background for next time */ - mlx5e_queue_update_stats(priv); + if (!mlx5e_monitor_counter_supported(priv)) { + /* update HW stats in background for next time */ + mlx5e_queue_update_stats(priv); + } if (mlx5e_is_uplink_rep(priv)) { stats->rx_packets = PPORT_802_3_GET(pstats, a_frames_received_ok); @@ -4901,6 +4904,8 @@ static void mlx5e_nic_enable(struct mlx5e_priv *priv) mlx5_lag_add(mdev, netdev); mlx5e_enable_async_events(priv); + if (mlx5e_monitor_counter_supported(priv)) + mlx5e_monitor_counter_init(priv); if (MLX5_ESWITCH_MANAGER(priv->mdev)) mlx5e_register_vport_reps(priv); @@ -4940,6 +4945,9 @@ static void mlx5e_nic_disable(struct mlx5e_priv *priv) if (MLX5_ESWITCH_MANAGER(priv->mdev)) mlx5e_unregister_vport_reps(priv); + if (mlx5e_monitor_counter_supported(priv)) + mlx5e_monitor_counter_cleanup(priv); + mlx5e_disable_async_events(priv); mlx5_lag_remove(mdev); } -- cgit v1.2.3-55-g7522 From 186599f89ea57e9b444dcb0793723b9148770161 Mon Sep 17 00:00:00 2001 From: YueHaibing Date: Wed, 12 Dec 2018 08:03:38 +0000 Subject: net/mlx5: Remove duplicated include from eswitch.c Remove duplicated include. Signed-off-by: YueHaibing Reviewed-by: Leon Romanovsky Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 1 - 1 file changed, 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index e6a9b19d8626..4f9dd0d4a081 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -39,7 +39,6 @@ #include "lib/eq.h" #include "eswitch.h" #include "fs_core.h" -#include "lib/eq.h" #define UPLINK_VPORT 0xFFFF -- cgit v1.2.3-55-g7522 From c784a28b025a22f8c791e14096fb7079236739c0 Mon Sep 17 00:00:00 2001 From: YueHaibing Date: Wed, 12 Dec 2018 08:33:53 +0000 Subject: net/mlx5e: Remove set but not used variable 'upriv' Fixes gcc '-Wunused-but-set-variable' warning: drivers/net/ethernet/mellanox/mlx5/core/en_rep.c: In function 'mlx5e_vport_rep_load': drivers/net/ethernet/mellanox/mlx5/core/en_rep.c:1490:21: warning: variable 'upriv' set but not used [-Wunused-but-set-variable] drivers/net/ethernet/mellanox/mlx5/core/en_rep.c: In function 'mlx5e_vport_rep_unload': drivers/net/ethernet/mellanox/mlx5/core/en_rep.c:1557:21: warning: variable 'upriv' set but not used [-Wunused-but-set-variable] It not used any more since commit ef381359e3a8 ("net/mlx5e: Replace egdev with indirect block notifications"). Also remove unused variable 'uplink_rpriv' after this change. Signed-off-by: YueHaibing Reviewed-by: Leon Romanovsky Acked-by: Saeed Mahameed Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 10 ---------- 1 file changed, 10 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index 85e51bd4147f..e5230049612f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1484,10 +1484,8 @@ mlx5e_nic_rep_unload(struct mlx5_eswitch_rep *rep) static int mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) { - struct mlx5e_rep_priv *uplink_rpriv; struct mlx5e_rep_priv *rpriv; struct net_device *netdev; - struct mlx5e_priv *upriv; int nch, err; rpriv = kzalloc(sizeof(*rpriv), GFP_KERNEL); @@ -1522,9 +1520,6 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) goto err_detach_netdev; } - uplink_rpriv = mlx5_eswitch_get_uplink_priv(dev->priv.eswitch, REP_ETH); - upriv = netdev_priv(uplink_rpriv->netdev); - err = register_netdev(netdev); if (err) { pr_warn("Failed to register representor netdev for vport %d\n", @@ -1552,14 +1547,9 @@ mlx5e_vport_rep_unload(struct mlx5_eswitch_rep *rep) struct mlx5e_rep_priv *rpriv = mlx5e_rep_to_rep_priv(rep); struct net_device *netdev = rpriv->netdev; struct mlx5e_priv *priv = netdev_priv(netdev); - struct mlx5e_rep_priv *uplink_rpriv; void *ppriv = priv->ppriv; - struct mlx5e_priv *upriv; unregister_netdev(netdev); - uplink_rpriv = mlx5_eswitch_get_uplink_priv(priv->mdev->priv.eswitch, - REP_ETH); - upriv = netdev_priv(uplink_rpriv->netdev); mlx5e_rep_neigh_cleanup(rpriv); mlx5e_detach_netdev(priv); mlx5e_destroy_netdev(priv); -- cgit v1.2.3-55-g7522 From 8252cf728ce230bab2f5898449ffe94ff02c3622 Mon Sep 17 00:00:00 2001 From: Roi Dayan Date: Wed, 12 Dec 2018 19:11:39 -0800 Subject: net/mlx5: Split the activate lag function into two routines Split the activate lag function in order to be symmetric with the deactivate lag function. Signed-off-by: Roi Dayan Signed-off-by: Shahar Klein Reviewed-by: Aviv Heller Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/lag.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c index 582b2f18010a..75f7c8efaab5 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c @@ -186,14 +186,12 @@ static void mlx5_infer_tx_affinity_mapping(struct lag_tracker *tracker, *port2 = 1; } -static void mlx5_activate_lag(struct mlx5_lag *ldev, - struct lag_tracker *tracker) +static int mlx5_create_lag(struct mlx5_lag *ldev, + struct lag_tracker *tracker) { struct mlx5_core_dev *dev0 = ldev->pf[0].dev; int err; - ldev->flags |= MLX5_LAG_FLAG_BONDED; - mlx5_infer_tx_affinity_mapping(tracker, &ldev->v2p_map[0], &ldev->v2p_map[1]); @@ -202,6 +200,14 @@ static void mlx5_activate_lag(struct mlx5_lag *ldev, mlx5_core_err(dev0, "Failed to create LAG (%d)\n", err); + return err; +} + +static void mlx5_activate_lag(struct mlx5_lag *ldev, + struct lag_tracker *tracker) +{ + ldev->flags |= MLX5_LAG_FLAG_BONDED; + mlx5_create_lag(ldev, tracker); } static void mlx5_deactivate_lag(struct mlx5_lag *ldev) -- cgit v1.2.3-55-g7522 From 3cfe432e1b3a6f5b48f146f72d6dde029a1640a1 Mon Sep 17 00:00:00 2001 From: Roi Dayan Date: Wed, 12 Dec 2018 19:11:40 -0800 Subject: net/mlx5: Add lag affinity info to log Report the initial LAG port affinity upon LAG creation. Signed-off-by: Roi Dayan Signed-off-by: Shahar Klein Reviewed-by: Aviv Heller Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/lag.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c index 75f7c8efaab5..448d9d5c689c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c @@ -195,6 +195,9 @@ static int mlx5_create_lag(struct mlx5_lag *ldev, mlx5_infer_tx_affinity_mapping(tracker, &ldev->v2p_map[0], &ldev->v2p_map[1]); + mlx5_core_info(dev0, "lag map port 1:%d port 2:%d", + ldev->v2p_map[0], ldev->v2p_map[1]); + err = mlx5_cmd_create_lag(dev0, ldev->v2p_map[0], ldev->v2p_map[1]); if (err) mlx5_core_err(dev0, -- cgit v1.2.3-55-g7522 From 4c283e6155988486e2c83a2f273f919ecd3ce721 Mon Sep 17 00:00:00 2001 From: Shahar Klein Date: Wed, 12 Dec 2018 19:11:41 -0800 Subject: net/mlx5: Fold the modify lag code into function Handle the code of modifying the lag affinity within a separate function. Signed-off-by: Shahar Klein Signed-off-by: Roi Dayan Reviewed-by: Aviv Heller Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/lag.c | 44 +++++++++++++++++---------- 1 file changed, 28 insertions(+), 16 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c index 448d9d5c689c..8c5c5e418d61 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c @@ -186,6 +186,32 @@ static void mlx5_infer_tx_affinity_mapping(struct lag_tracker *tracker, *port2 = 1; } +static void mlx5_modify_lag(struct mlx5_lag *ldev, + struct lag_tracker *tracker) +{ + struct mlx5_core_dev *dev0 = ldev->pf[0].dev; + u8 v2p_port1, v2p_port2; + int err; + + mlx5_infer_tx_affinity_mapping(tracker, &v2p_port1, + &v2p_port2); + + if (v2p_port1 != ldev->v2p_map[0] || + v2p_port2 != ldev->v2p_map[1]) { + ldev->v2p_map[0] = v2p_port1; + ldev->v2p_map[1] = v2p_port2; + + mlx5_core_info(dev0, "modify lag map port 1:%d port 2:%d", + ldev->v2p_map[0], ldev->v2p_map[1]); + + err = mlx5_cmd_modify_lag(dev0, v2p_port1, v2p_port2); + if (err) + mlx5_core_err(dev0, + "Failed to modify LAG (%d)\n", + err); + } +} + static int mlx5_create_lag(struct mlx5_lag *ldev, struct lag_tracker *tracker) { @@ -232,8 +258,7 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) struct mlx5_core_dev *dev0 = ldev->pf[0].dev; struct mlx5_core_dev *dev1 = ldev->pf[1].dev; struct lag_tracker tracker; - u8 v2p_port1, v2p_port2; - int i, err; + int i; bool do_bond; if (!dev0 || !dev1) @@ -255,20 +280,7 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) mlx5_add_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); mlx5_nic_vport_enable_roce(dev1); } else if (do_bond && mlx5_lag_is_bonded(ldev)) { - mlx5_infer_tx_affinity_mapping(&tracker, &v2p_port1, - &v2p_port2); - - if ((v2p_port1 != ldev->v2p_map[0]) || - (v2p_port2 != ldev->v2p_map[1])) { - ldev->v2p_map[0] = v2p_port1; - ldev->v2p_map[1] = v2p_port2; - - err = mlx5_cmd_modify_lag(dev0, v2p_port1, v2p_port2); - if (err) - mlx5_core_err(dev0, - "Failed to modify LAG (%d)\n", - err); - } + mlx5_modify_lag(ldev, &tracker); } else if (!do_bond && mlx5_lag_is_bonded(ldev)) { mlx5_remove_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); mlx5_nic_vport_disable_roce(dev1); -- cgit v1.2.3-55-g7522 From fadd59fc50d010145f251db583c7ccef37393d19 Mon Sep 17 00:00:00 2001 From: Aviv Heller Date: Tue, 4 Dec 2018 21:24:46 +0200 Subject: net/mlx5: Introduce inter-device communication mechanism This introduces devcom, a generic mechanism for performing operations on both physical functions of the same Connect-X card. The first user of this API is merged eswitch, which will be introduced in subsequent patches. Signed-off-by: Aviv Heller Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/Makefile | 2 +- .../net/ethernet/mellanox/mlx5/core/lib/devcom.c | 255 +++++++++++++++++++++ .../net/ethernet/mellanox/mlx5/core/lib/devcom.h | 44 ++++ drivers/net/ethernet/mellanox/mlx5/core/main.c | 14 +- include/linux/mlx5/driver.h | 2 + 5 files changed, 313 insertions(+), 4 deletions(-) create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.c create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile index 9678051b8ff1..9de9abacf7f6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile +++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile @@ -15,7 +15,7 @@ mlx5_core-y := main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \ health.o mcg.o cq.o alloc.o qp.o port.o mr.o pd.o \ mad.o transobj.o vport.o sriov.o fs_cmd.o fs_core.o \ fs_counters.o rl.o lag.o dev.o events.o wq.o lib/gid.o \ - diag/fs_tracepoint.o diag/fw_tracer.o + lib/devcom.o diag/fs_tracepoint.o diag/fw_tracer.o # # Netdev basic diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.c new file mode 100644 index 000000000000..bced2efe9bef --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.c @@ -0,0 +1,255 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB +/* Copyright (c) 2018 Mellanox Technologies */ + +#include +#include "lib/devcom.h" + +static LIST_HEAD(devcom_list); + +#define devcom_for_each_component(priv, comp, iter) \ + for (iter = 0; \ + comp = &(priv)->components[iter], iter < MLX5_DEVCOM_NUM_COMPONENTS; \ + iter++) + +struct mlx5_devcom_component { + struct { + void *data; + } device[MLX5_MAX_PORTS]; + + mlx5_devcom_event_handler_t handler; + struct rw_semaphore sem; + bool paired; +}; + +struct mlx5_devcom_list { + struct list_head list; + + struct mlx5_devcom_component components[MLX5_DEVCOM_NUM_COMPONENTS]; + struct mlx5_core_dev *devs[MLX5_MAX_PORTS]; +}; + +struct mlx5_devcom { + struct mlx5_devcom_list *priv; + int idx; +}; + +static struct mlx5_devcom_list *mlx5_devcom_list_alloc(void) +{ + struct mlx5_devcom_component *comp; + struct mlx5_devcom_list *priv; + int i; + + priv = kzalloc(sizeof(*priv), GFP_KERNEL); + if (!priv) + return NULL; + + devcom_for_each_component(priv, comp, i) + init_rwsem(&comp->sem); + + return priv; +} + +static struct mlx5_devcom *mlx5_devcom_alloc(struct mlx5_devcom_list *priv, + u8 idx) +{ + struct mlx5_devcom *devcom; + + devcom = kzalloc(sizeof(*devcom), GFP_KERNEL); + if (!devcom) + return NULL; + + devcom->priv = priv; + devcom->idx = idx; + return devcom; +} + +/* Must be called with intf_mutex held */ +struct mlx5_devcom *mlx5_devcom_register_device(struct mlx5_core_dev *dev) +{ + struct mlx5_devcom_list *priv = NULL, *iter; + struct mlx5_devcom *devcom = NULL; + bool new_priv = false; + u64 sguid0, sguid1; + int idx, i; + + if (!mlx5_core_is_pf(dev)) + return NULL; + + sguid0 = mlx5_query_nic_system_image_guid(dev); + list_for_each_entry(iter, &devcom_list, list) { + struct mlx5_core_dev *tmp_dev = NULL; + + idx = -1; + for (i = 0; i < MLX5_MAX_PORTS; i++) { + if (iter->devs[i]) + tmp_dev = iter->devs[i]; + else + idx = i; + } + + if (idx == -1) + continue; + + sguid1 = mlx5_query_nic_system_image_guid(tmp_dev); + if (sguid0 != sguid1) + continue; + + priv = iter; + break; + } + + if (!priv) { + priv = mlx5_devcom_list_alloc(); + if (!priv) + return ERR_PTR(-ENOMEM); + + idx = 0; + new_priv = true; + } + + priv->devs[idx] = dev; + devcom = mlx5_devcom_alloc(priv, idx); + if (!devcom) { + kfree(priv); + return ERR_PTR(-ENOMEM); + } + + if (new_priv) + list_add(&priv->list, &devcom_list); + + return devcom; +} + +/* Must be called with intf_mutex held */ +void mlx5_devcom_unregister_device(struct mlx5_devcom *devcom) +{ + struct mlx5_devcom_list *priv; + int i; + + if (IS_ERR_OR_NULL(devcom)) + return; + + priv = devcom->priv; + priv->devs[devcom->idx] = NULL; + + kfree(devcom); + + for (i = 0; i < MLX5_MAX_PORTS; i++) + if (priv->devs[i]) + break; + + if (i != MLX5_MAX_PORTS) + return; + + list_del(&priv->list); + kfree(priv); +} + +void mlx5_devcom_register_component(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id, + mlx5_devcom_event_handler_t handler, + void *data) +{ + struct mlx5_devcom_component *comp; + + if (IS_ERR_OR_NULL(devcom)) + return; + + WARN_ON(!data); + + comp = &devcom->priv->components[id]; + down_write(&comp->sem); + comp->handler = handler; + comp->device[devcom->idx].data = data; + up_write(&comp->sem); +} + +void mlx5_devcom_unregister_component(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id) +{ + struct mlx5_devcom_component *comp; + + if (IS_ERR_OR_NULL(devcom)) + return; + + comp = &devcom->priv->components[id]; + down_write(&comp->sem); + comp->device[devcom->idx].data = NULL; + up_write(&comp->sem); +} + +int mlx5_devcom_send_event(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id, + int event, + void *event_data) +{ + struct mlx5_devcom_component *comp; + int err = -ENODEV, i; + + if (IS_ERR_OR_NULL(devcom)) + return err; + + comp = &devcom->priv->components[id]; + down_write(&comp->sem); + for (i = 0; i < MLX5_MAX_PORTS; i++) + if (i != devcom->idx && comp->device[i].data) { + err = comp->handler(event, comp->device[i].data, + event_data); + break; + } + + up_write(&comp->sem); + return err; +} + +void mlx5_devcom_set_paired(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id, + bool paired) +{ + struct mlx5_devcom_component *comp; + + comp = &devcom->priv->components[id]; + WARN_ON(!rwsem_is_locked(&comp->sem)); + + comp->paired = paired; +} + +bool mlx5_devcom_is_paired(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id) +{ + if (IS_ERR_OR_NULL(devcom)) + return false; + + return devcom->priv->components[id].paired; +} + +void *mlx5_devcom_get_peer_data(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id) +{ + struct mlx5_devcom_component *comp; + int i; + + if (IS_ERR_OR_NULL(devcom)) + return NULL; + + comp = &devcom->priv->components[id]; + down_read(&comp->sem); + if (!comp->paired) { + up_read(&comp->sem); + return NULL; + } + + for (i = 0; i < MLX5_MAX_PORTS; i++) + if (i != devcom->idx) + break; + + return comp->device[i].data; +} + +void mlx5_devcom_release_peer_data(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id) +{ + struct mlx5_devcom_component *comp = &devcom->priv->components[id]; + + up_read(&comp->sem); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h new file mode 100644 index 000000000000..f2d338b187a6 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h @@ -0,0 +1,44 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* Copyright (c) 2018 Mellanox Technologies */ + +#ifndef __LIB_MLX5_DEVCOM_H__ +#define __LIB_MLX5_DEVCOM_H__ + +#include + +enum mlx5_devcom_components { + MLX5_DEVCOM_NUM_COMPONENTS, +}; + +typedef int (*mlx5_devcom_event_handler_t)(int event, + void *my_data, + void *event_data); + +struct mlx5_devcom *mlx5_devcom_register_device(struct mlx5_core_dev *dev); +void mlx5_devcom_unregister_device(struct mlx5_devcom *devcom); + +void mlx5_devcom_register_component(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id, + mlx5_devcom_event_handler_t handler, + void *data); +void mlx5_devcom_unregister_component(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id); + +int mlx5_devcom_send_event(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id, + int event, + void *event_data); + +void mlx5_devcom_set_paired(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id, + bool paired); +bool mlx5_devcom_is_paired(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id); + +void *mlx5_devcom_get_peer_data(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id); +void mlx5_devcom_release_peer_data(struct mlx5_devcom *devcom, + enum mlx5_devcom_components id); + +#endif + diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 778995573812..c23553164e0d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -63,6 +63,7 @@ #include "accel/tls.h" #include "lib/clock.h" #include "lib/vxlan.h" +#include "lib/devcom.h" #include "diag/fw_tracer.h" MODULE_AUTHOR("Eli Cohen "); @@ -722,16 +723,21 @@ static int mlx5_init_once(struct mlx5_core_dev *dev, struct mlx5_priv *priv) struct pci_dev *pdev = dev->pdev; int err; + priv->devcom = mlx5_devcom_register_device(dev); + if (IS_ERR(priv->devcom)) + dev_err(&pdev->dev, "failed to register with devcom (0x%p)\n", + priv->devcom); + err = mlx5_query_board_id(dev); if (err) { dev_err(&pdev->dev, "query board id failed\n"); - goto out; + goto err_devcom; } err = mlx5_eq_table_init(dev); if (err) { dev_err(&pdev->dev, "failed to initialize eq\n"); - goto out; + goto err_devcom; } err = mlx5_events_init(dev); @@ -807,8 +813,9 @@ err_events_cleanup: mlx5_events_cleanup(dev); err_eq_cleanup: mlx5_eq_table_cleanup(dev); +err_devcom: + mlx5_devcom_unregister_device(dev->priv.devcom); -out: return err; } @@ -828,6 +835,7 @@ static void mlx5_cleanup_once(struct mlx5_core_dev *dev) mlx5_cq_debugfs_cleanup(dev); mlx5_events_cleanup(dev); mlx5_eq_table_cleanup(dev); + mlx5_devcom_unregister_device(dev->priv.devcom); } static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index cc29e880c733..cd7af5d0311b 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -486,6 +486,7 @@ struct mlx5_events; struct mlx5_mpfs; struct mlx5_eswitch; struct mlx5_lag; +struct mlx5_devcom; struct mlx5_eq_table; struct mlx5_rate_limit { @@ -560,6 +561,7 @@ struct mlx5_priv { struct mlx5_eswitch *eswitch; struct mlx5_core_sriov sriov; struct mlx5_lag *lag; + struct mlx5_devcom *devcom; unsigned long pci_dev_data; struct mlx5_fc_stats fc_stats; struct mlx5_rl_table rl_table; -- cgit v1.2.3-55-g7522 From ac004b8321281b9da7d9a99674241e60e49ffc8a Mon Sep 17 00:00:00 2001 From: Roi Dayan Date: Sun, 11 Nov 2018 17:50:16 +0200 Subject: net/mlx5e: E-Switch, Add peer miss rules In the sriov offloads mode, packets that are not matched by any other rule are sent towards the e-switch vport manager for further processing. Under upcoming patches (e.g for uplink LAG), packets sent from VF vports belonging to esw0 (e-switch related to PF0) might end up in esw1 (e-switch related to PF1) due to muxing logic applied by the FW. In such a case we still want the missed packet to be sent to the "base" esw manager vport in order to present the control plane a consistent view of the source (VF reresentor) port. Signed-off-by: Roi Dayan Signed-off-by: Aviv Heller Signed-off-by: Shahar Klein Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 2 + .../ethernet/mellanox/mlx5/core/eswitch_offloads.c | 222 ++++++++++++++++++++- .../net/ethernet/mellanox/mlx5/core/lib/devcom.h | 2 + 3 files changed, 225 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h index 87c9dea9bccf..9dba6ad5744d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h @@ -143,6 +143,8 @@ struct mlx5_eswitch_fdb { struct offloads_fdb { struct mlx5_flow_table *slow_fdb; struct mlx5_flow_group *send_to_vport_grp; + struct mlx5_flow_group *peer_miss_grp; + struct mlx5_flow_handle **peer_miss_rules; struct mlx5_flow_group *miss_grp; struct mlx5_flow_handle *miss_rule_uni; struct mlx5_flow_handle *miss_rule_multi; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index bde1fb8c284b..a6927ca3d4ca 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -39,6 +39,7 @@ #include "eswitch.h" #include "en.h" #include "fs_core.h" +#include "lib/devcom.h" enum { FDB_FAST_PATH = 0, @@ -541,6 +542,98 @@ void mlx5_eswitch_del_send_to_vport_rule(struct mlx5_flow_handle *rule) mlx5_del_flow_rules(rule); } +static void peer_miss_rules_setup(struct mlx5_core_dev *peer_dev, + struct mlx5_flow_spec *spec, + struct mlx5_flow_destination *dest) +{ + void *misc = MLX5_ADDR_OF(fte_match_param, spec->match_value, + misc_parameters); + + MLX5_SET(fte_match_set_misc, misc, source_eswitch_owner_vhca_id, + MLX5_CAP_GEN(peer_dev, vhca_id)); + + spec->match_criteria_enable = MLX5_MATCH_MISC_PARAMETERS; + + misc = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, + misc_parameters); + MLX5_SET_TO_ONES(fte_match_set_misc, misc, source_port); + MLX5_SET_TO_ONES(fte_match_set_misc, misc, + source_eswitch_owner_vhca_id); + + dest->type = MLX5_FLOW_DESTINATION_TYPE_VPORT; + dest->vport.num = 0; + dest->vport.vhca_id = MLX5_CAP_GEN(peer_dev, vhca_id); + dest->vport.vhca_id_valid = 1; +} + +static int esw_add_fdb_peer_miss_rules(struct mlx5_eswitch *esw, + struct mlx5_core_dev *peer_dev) +{ + struct mlx5_flow_destination dest = {}; + struct mlx5_flow_act flow_act = {0}; + struct mlx5_flow_handle **flows; + struct mlx5_flow_handle *flow; + struct mlx5_flow_spec *spec; + /* total vports is the same for both e-switches */ + int nvports = esw->total_vports; + void *misc; + int err, i; + + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); + if (!spec) + return -ENOMEM; + + peer_miss_rules_setup(peer_dev, spec, &dest); + + flows = kvzalloc(nvports * sizeof(*flows), GFP_KERNEL); + if (!flows) { + err = -ENOMEM; + goto alloc_flows_err; + } + + flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST; + misc = MLX5_ADDR_OF(fte_match_param, spec->match_value, + misc_parameters); + + for (i = 1; i < nvports; i++) { + MLX5_SET(fte_match_set_misc, misc, source_port, i); + flow = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb, + spec, &flow_act, &dest, 1); + if (IS_ERR(flow)) { + err = PTR_ERR(flow); + esw_warn(esw->dev, "FDB: Failed to add peer miss flow rule err %d\n", err); + goto add_flow_err; + } + flows[i] = flow; + } + + esw->fdb_table.offloads.peer_miss_rules = flows; + + kvfree(spec); + return 0; + +add_flow_err: + for (i--; i > 0; i--) + mlx5_del_flow_rules(flows[i]); + kvfree(flows); +alloc_flows_err: + kvfree(spec); + return err; +} + +static void esw_del_fdb_peer_miss_rules(struct mlx5_eswitch *esw) +{ + struct mlx5_flow_handle **flows; + int i; + + flows = esw->fdb_table.offloads.peer_miss_rules; + + for (i = 1; i < esw->total_vports; i++) + mlx5_del_flow_rules(flows[i]); + + kvfree(flows); +} + static int esw_add_fdb_miss_rule(struct mlx5_eswitch *esw) { struct mlx5_flow_act flow_act = {0}; @@ -811,7 +904,8 @@ static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw, int nvports) esw->fdb_table.offloads.fdb_left[i] = ESW_POOLS[i] <= fdb_max ? ESW_SIZE / ESW_POOLS[i] : 0; - table_size = nvports * MAX_SQ_NVPORTS + MAX_PF_SQ + 2; + table_size = nvports * MAX_SQ_NVPORTS + MAX_PF_SQ + 2 + + esw->total_vports; /* create the slow path fdb with encap set, so further table instances * can be created at run time while VFs are probed if the FW allows that. @@ -866,6 +960,34 @@ static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw, int nvports) } esw->fdb_table.offloads.send_to_vport_grp = g; + /* create peer esw miss group */ + memset(flow_group_in, 0, inlen); + MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable, + MLX5_MATCH_MISC_PARAMETERS); + + match_criteria = MLX5_ADDR_OF(create_flow_group_in, flow_group_in, + match_criteria); + + MLX5_SET_TO_ONES(fte_match_param, match_criteria, + misc_parameters.source_port); + MLX5_SET_TO_ONES(fte_match_param, match_criteria, + misc_parameters.source_eswitch_owner_vhca_id); + + MLX5_SET(create_flow_group_in, flow_group_in, + source_eswitch_owner_vhca_id_valid, 1); + MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, ix); + MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, + ix + esw->total_vports - 1); + ix += esw->total_vports; + + g = mlx5_create_flow_group(fdb, flow_group_in); + if (IS_ERR(g)) { + err = PTR_ERR(g); + esw_warn(dev, "Failed to create peer miss flow group err(%d)\n", err); + goto peer_miss_err; + } + esw->fdb_table.offloads.peer_miss_grp = g; + /* create miss group */ memset(flow_group_in, 0, inlen); MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable, @@ -898,6 +1020,8 @@ static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw, int nvports) miss_rule_err: mlx5_destroy_flow_group(esw->fdb_table.offloads.miss_grp); miss_err: + mlx5_destroy_flow_group(esw->fdb_table.offloads.peer_miss_grp); +peer_miss_err: mlx5_destroy_flow_group(esw->fdb_table.offloads.send_to_vport_grp); send_vport_err: esw_destroy_offloads_fast_fdb_tables(esw); @@ -917,6 +1041,7 @@ static void esw_destroy_offloads_fdb_tables(struct mlx5_eswitch *esw) mlx5_del_flow_rules(esw->fdb_table.offloads.miss_rule_multi); mlx5_del_flow_rules(esw->fdb_table.offloads.miss_rule_uni); mlx5_destroy_flow_group(esw->fdb_table.offloads.send_to_vport_grp); + mlx5_destroy_flow_group(esw->fdb_table.offloads.peer_miss_grp); mlx5_destroy_flow_group(esw->fdb_table.offloads.miss_grp); mlx5_destroy_flow_table(esw->fdb_table.offloads.slow_fdb); @@ -1173,6 +1298,99 @@ err_reps: return err; } +#define ESW_OFFLOADS_DEVCOM_PAIR (0) +#define ESW_OFFLOADS_DEVCOM_UNPAIR (1) + +static int mlx5_esw_offloads_pair(struct mlx5_eswitch *esw, + struct mlx5_eswitch *peer_esw) +{ + int err; + + err = esw_add_fdb_peer_miss_rules(esw, peer_esw->dev); + if (err) + return err; + + return 0; +} + +static void mlx5_esw_offloads_unpair(struct mlx5_eswitch *esw) +{ + esw_del_fdb_peer_miss_rules(esw); +} + +static int mlx5_esw_offloads_devcom_event(int event, + void *my_data, + void *event_data) +{ + struct mlx5_eswitch *esw = my_data; + struct mlx5_eswitch *peer_esw = event_data; + struct mlx5_devcom *devcom = esw->dev->priv.devcom; + int err; + + switch (event) { + case ESW_OFFLOADS_DEVCOM_PAIR: + err = mlx5_esw_offloads_pair(esw, peer_esw); + if (err) + goto err_out; + + err = mlx5_esw_offloads_pair(peer_esw, esw); + if (err) + goto err_pair; + + mlx5_devcom_set_paired(devcom, MLX5_DEVCOM_ESW_OFFLOADS, true); + break; + + case ESW_OFFLOADS_DEVCOM_UNPAIR: + if (!mlx5_devcom_is_paired(devcom, MLX5_DEVCOM_ESW_OFFLOADS)) + break; + + mlx5_devcom_set_paired(devcom, MLX5_DEVCOM_ESW_OFFLOADS, false); + mlx5_esw_offloads_unpair(peer_esw); + mlx5_esw_offloads_unpair(esw); + break; + } + + return 0; + +err_pair: + mlx5_esw_offloads_unpair(esw); + +err_out: + mlx5_core_err(esw->dev, "esw offloads devcom event failure, event %u err %d", + event, err); + return err; +} + +static void esw_offloads_devcom_init(struct mlx5_eswitch *esw) +{ + struct mlx5_devcom *devcom = esw->dev->priv.devcom; + + if (!MLX5_CAP_ESW(esw->dev, merged_eswitch)) + return; + + mlx5_devcom_register_component(devcom, + MLX5_DEVCOM_ESW_OFFLOADS, + mlx5_esw_offloads_devcom_event, + esw); + + mlx5_devcom_send_event(devcom, + MLX5_DEVCOM_ESW_OFFLOADS, + ESW_OFFLOADS_DEVCOM_PAIR, esw); +} + +static void esw_offloads_devcom_cleanup(struct mlx5_eswitch *esw) +{ + struct mlx5_devcom *devcom = esw->dev->priv.devcom; + + if (!MLX5_CAP_ESW(esw->dev, merged_eswitch)) + return; + + mlx5_devcom_send_event(devcom, MLX5_DEVCOM_ESW_OFFLOADS, + ESW_OFFLOADS_DEVCOM_UNPAIR, esw); + + mlx5_devcom_unregister_component(devcom, MLX5_DEVCOM_ESW_OFFLOADS); +} + int esw_offloads_init(struct mlx5_eswitch *esw, int nvports) { int err; @@ -1195,6 +1413,7 @@ int esw_offloads_init(struct mlx5_eswitch *esw, int nvports) if (err) goto err_reps; + esw_offloads_devcom_init(esw); return 0; err_reps: @@ -1233,6 +1452,7 @@ static int esw_offloads_stop(struct mlx5_eswitch *esw, void esw_offloads_cleanup(struct mlx5_eswitch *esw, int nvports) { + esw_offloads_devcom_cleanup(esw); esw_offloads_unload_reps(esw, nvports); esw_destroy_vport_rx_group(esw); esw_destroy_offloads_table(esw); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h index f2d338b187a6..939d5bf1581b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h @@ -7,6 +7,8 @@ #include enum mlx5_devcom_components { + MLX5_DEVCOM_ESW_OFFLOADS, + MLX5_DEVCOM_NUM_COMPONENTS, }; -- cgit v1.2.3-55-g7522 From 04de7dda7394fa9c2b0fc9cec65661d9b4f0d04d Mon Sep 17 00:00:00 2001 From: Roi Dayan Date: Sun, 11 Nov 2018 22:24:03 +0200 Subject: net/mlx5e: Infrastructure for duplicated offloading of TC flows Under uplink LAG or multipath schemes, traffic that matches one flow might arrive on both uplink ports and transmitted through both as part of supporting aggregation and high-availability. To cope with the fact that the SW model might use logical SW port (e.g uplink team or bond) but we have two HW ports with e-switch on each, there are cases where in order to offload a SW TC rule we need to duplicate it to two HW flows. Since each HW rule has its own counter we also aggregate the counter of both rules when a flow stats query is executed from user-space. Introduce the changes for the different elements (add/delete/stats), currently nothing is duplicated. Signed-off-by: Roi Dayan Signed-off-by: Aviv Heller Signed-off-by: Shahar Klein Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 174 ++++++++++++++++++++- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 2 + .../ethernet/mellanox/mlx5/core/eswitch_offloads.c | 8 +- 3 files changed, 176 insertions(+), 8 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 0921213561cb..eacccac05dda 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -52,6 +52,7 @@ #include "fs_core.h" #include "en/port.h" #include "en/tc_tun.h" +#include "lib/devcom.h" struct mlx5_nic_flow_attr { u32 action; @@ -74,6 +75,7 @@ enum { MLX5E_TC_FLOW_HAIRPIN = BIT(MLX5E_TC_FLOW_BASE + 3), MLX5E_TC_FLOW_HAIRPIN_RSS = BIT(MLX5E_TC_FLOW_BASE + 4), MLX5E_TC_FLOW_SLOW = BIT(MLX5E_TC_FLOW_BASE + 5), + MLX5E_TC_FLOW_DUP = BIT(MLX5E_TC_FLOW_BASE + 6), }; #define MLX5E_TC_MAX_SPLITS 1 @@ -111,8 +113,10 @@ struct mlx5e_tc_flow { * destinations. */ struct encap_flow_item encaps[MLX5_MAX_FLOW_FWD_VPORTS]; + struct mlx5e_tc_flow *peer_flow; struct list_head mod_hdr; /* flows sharing the same mod hdr ID */ struct list_head hairpin; /* flows sharing the same hairpin */ + struct list_head peer; /* flows with peer flow */ union { struct mlx5_esw_flow_attr esw_attr[0]; struct mlx5_nic_flow_attr nic_attr[0]; @@ -1249,13 +1253,48 @@ static void mlx5e_detach_encap(struct mlx5e_priv *priv, } } +static void __mlx5e_tc_del_fdb_peer_flow(struct mlx5e_tc_flow *flow) +{ + struct mlx5_eswitch *esw = flow->priv->mdev->priv.eswitch; + + if (!(flow->flags & MLX5E_TC_FLOW_ESWITCH) || + !(flow->flags & MLX5E_TC_FLOW_DUP)) + return; + + mutex_lock(&esw->offloads.peer_mutex); + list_del(&flow->peer); + mutex_unlock(&esw->offloads.peer_mutex); + + flow->flags &= ~MLX5E_TC_FLOW_DUP; + + mlx5e_tc_del_fdb_flow(flow->peer_flow->priv, flow->peer_flow); + kvfree(flow->peer_flow); + flow->peer_flow = NULL; +} + +static void mlx5e_tc_del_fdb_peer_flow(struct mlx5e_tc_flow *flow) +{ + struct mlx5_core_dev *dev = flow->priv->mdev; + struct mlx5_devcom *devcom = dev->priv.devcom; + struct mlx5_eswitch *peer_esw; + + peer_esw = mlx5_devcom_get_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS); + if (!peer_esw) + return; + + __mlx5e_tc_del_fdb_peer_flow(flow); + mlx5_devcom_release_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS); +} + static void mlx5e_tc_del_flow(struct mlx5e_priv *priv, struct mlx5e_tc_flow *flow) { - if (flow->flags & MLX5E_TC_FLOW_ESWITCH) + if (flow->flags & MLX5E_TC_FLOW_ESWITCH) { + mlx5e_tc_del_fdb_peer_flow(flow); mlx5e_tc_del_fdb_flow(priv, flow); - else + } else { mlx5e_tc_del_nic_flow(priv, flow); + } } @@ -2660,6 +2699,11 @@ static struct rhashtable *get_tc_ht(struct mlx5e_priv *priv) return &priv->fs.tc.ht; } +static bool is_peer_flow_needed(struct mlx5e_tc_flow *flow) +{ + return false; +} + static int mlx5e_alloc_flow(struct mlx5e_priv *priv, int attr_size, struct tc_cls_flower_offload *f, u16 flow_flags, @@ -2693,11 +2737,13 @@ err_free: } static int -mlx5e_add_fdb_flow(struct mlx5e_priv *priv, - struct tc_cls_flower_offload *f, - u16 flow_flags, - struct net_device *filter_dev, - struct mlx5e_tc_flow **__flow) +__mlx5e_add_fdb_flow(struct mlx5e_priv *priv, + struct tc_cls_flower_offload *f, + u16 flow_flags, + struct net_device *filter_dev, + struct mlx5_eswitch_rep *in_rep, + struct mlx5_core_dev *in_mdev, + struct mlx5e_tc_flow **__flow) { struct netlink_ext_ack *extack = f->common.extack; struct mlx5e_tc_flow_parse_attr *parse_attr; @@ -2723,6 +2769,8 @@ mlx5e_add_fdb_flow(struct mlx5e_priv *priv, if (err) goto err_free; + flow->esw_attr->in_rep = in_rep; + flow->esw_attr->in_mdev = in_mdev; err = mlx5e_tc_add_fdb_flow(priv, parse_attr, flow, extack); if (err) goto err_free; @@ -2738,6 +2786,87 @@ out: return err; } +static int mlx5e_tc_add_fdb_peer_flow(struct tc_cls_flower_offload *f, + struct mlx5e_tc_flow *flow) +{ + struct mlx5e_priv *priv = flow->priv, *peer_priv; + struct mlx5_eswitch *esw = priv->mdev->priv.eswitch, *peer_esw; + struct mlx5_devcom *devcom = priv->mdev->priv.devcom; + struct mlx5e_tc_flow_parse_attr *parse_attr; + struct mlx5e_rep_priv *peer_urpriv; + struct mlx5e_tc_flow *peer_flow; + struct mlx5_core_dev *in_mdev; + int err = 0; + + peer_esw = mlx5_devcom_get_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS); + if (!peer_esw) + return -ENODEV; + + peer_urpriv = mlx5_eswitch_get_uplink_priv(peer_esw, REP_ETH); + peer_priv = netdev_priv(peer_urpriv->netdev); + + /* in_mdev is assigned of which the packet originated from. + * So packets redirected to uplink use the same mdev of the + * original flow and packets redirected from uplink use the + * peer mdev. + */ + if (flow->esw_attr->in_rep->vport == FDB_UPLINK_VPORT) + in_mdev = peer_priv->mdev; + else + in_mdev = priv->mdev; + + parse_attr = flow->esw_attr->parse_attr; + err = __mlx5e_add_fdb_flow(peer_priv, f, flow->flags, + parse_attr->filter_dev, + flow->esw_attr->in_rep, in_mdev, &peer_flow); + if (err) + goto out; + + flow->peer_flow = peer_flow; + flow->flags |= MLX5E_TC_FLOW_DUP; + mutex_lock(&esw->offloads.peer_mutex); + list_add_tail(&flow->peer, &esw->offloads.peer_flows); + mutex_unlock(&esw->offloads.peer_mutex); + +out: + mlx5_devcom_release_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS); + return err; +} + +static int +mlx5e_add_fdb_flow(struct mlx5e_priv *priv, + struct tc_cls_flower_offload *f, + u16 flow_flags, + struct net_device *filter_dev, + struct mlx5e_tc_flow **__flow) +{ + struct mlx5e_rep_priv *rpriv = priv->ppriv; + struct mlx5_eswitch_rep *in_rep = rpriv->rep; + struct mlx5_core_dev *in_mdev = priv->mdev; + struct mlx5e_tc_flow *flow; + int err; + + err = __mlx5e_add_fdb_flow(priv, f, flow_flags, filter_dev, in_rep, + in_mdev, &flow); + if (err) + goto out; + + if (is_peer_flow_needed(flow)) { + err = mlx5e_tc_add_fdb_peer_flow(f, flow); + if (err) { + mlx5e_tc_del_fdb_flow(priv, flow); + goto out; + } + } + + *__flow = flow; + + return 0; + +out: + return err; +} + static int mlx5e_add_nic_flow(struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, @@ -2882,7 +3011,9 @@ int mlx5e_delete_flower(struct net_device *dev, struct mlx5e_priv *priv, int mlx5e_stats_flower(struct net_device *dev, struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, int flags) { + struct mlx5_devcom *devcom = priv->mdev->priv.devcom; struct rhashtable *tc_ht = get_tc_ht(priv); + struct mlx5_eswitch *peer_esw; struct mlx5e_tc_flow *flow; struct mlx5_fc *counter; u64 bytes; @@ -2902,6 +3033,27 @@ int mlx5e_stats_flower(struct net_device *dev, struct mlx5e_priv *priv, mlx5_fc_query_cached(counter, &bytes, &packets, &lastuse); + peer_esw = mlx5_devcom_get_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS); + if (!peer_esw) + goto out; + + if ((flow->flags & MLX5E_TC_FLOW_DUP) && + (flow->peer_flow->flags & MLX5E_TC_FLOW_OFFLOADED)) { + u64 bytes2; + u64 packets2; + u64 lastuse2; + + counter = mlx5e_tc_get_counter(flow->peer_flow); + mlx5_fc_query_cached(counter, &bytes2, &packets2, &lastuse2); + + bytes += bytes2; + packets += packets2; + lastuse = max_t(u64, lastuse, lastuse2); + } + + mlx5_devcom_release_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS); + +out: tcf_exts_stats_update(f->exts, bytes, packets, lastuse); return 0; @@ -3014,3 +3166,11 @@ int mlx5e_tc_num_filters(struct mlx5e_priv *priv) return atomic_read(&tc_ht->nelems); } + +void mlx5e_tc_clean_fdb_peer_flows(struct mlx5_eswitch *esw) +{ + struct mlx5e_tc_flow *flow, *tmp; + + list_for_each_entry_safe(flow, tmp, &esw->offloads.peer_flows, peer) + __mlx5e_tc_del_fdb_peer_flow(flow); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h index 9dba6ad5744d..4d048f7e703b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h @@ -167,6 +167,8 @@ struct mlx5_esw_offload { struct mlx5_flow_table *ft_offloads; struct mlx5_flow_group *vport_rx_group; struct mlx5_eswitch_rep *vport_reps; + struct list_head peer_flows; + struct mutex peer_mutex; DECLARE_HASHTABLE(encap_tbl, 8); DECLARE_HASHTABLE(mod_hdr_tbl, 8); u8 inline_mode; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index a6927ca3d4ca..76cb57202474 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -563,7 +563,7 @@ static void peer_miss_rules_setup(struct mlx5_core_dev *peer_dev, dest->type = MLX5_FLOW_DESTINATION_TYPE_VPORT; dest->vport.num = 0; dest->vport.vhca_id = MLX5_CAP_GEN(peer_dev, vhca_id); - dest->vport.vhca_id_valid = 1; + dest->vport.flags |= MLX5_FLOW_DEST_VPORT_VHCA_ID; } static int esw_add_fdb_peer_miss_rules(struct mlx5_eswitch *esw, @@ -1313,8 +1313,11 @@ static int mlx5_esw_offloads_pair(struct mlx5_eswitch *esw, return 0; } +void mlx5e_tc_clean_fdb_peer_flows(struct mlx5_eswitch *esw); + static void mlx5_esw_offloads_unpair(struct mlx5_eswitch *esw) { + mlx5e_tc_clean_fdb_peer_flows(esw); esw_del_fdb_peer_miss_rules(esw); } @@ -1365,6 +1368,9 @@ static void esw_offloads_devcom_init(struct mlx5_eswitch *esw) { struct mlx5_devcom *devcom = esw->dev->priv.devcom; + INIT_LIST_HEAD(&esw->offloads.peer_flows); + mutex_init(&esw->offloads.peer_mutex); + if (!MLX5_CAP_ESW(esw->dev, merged_eswitch)) return; -- cgit v1.2.3-55-g7522 From f9392795e2e35449a7dca46574b1a86eace20e9c Mon Sep 17 00:00:00 2001 From: Shahar Klein Date: Thu, 30 Nov 2017 17:55:46 +0200 Subject: net/mlx5e: Enhance flow counter scheme for offloaded TC eswitch rules Assign a counter dev attribute according to device capability and use it for management of counters related to offloaded eswitch TC flows. With upcoming support for uplink LAG, we have two HW rules per one logical SW (TC) rule. Although the HW supports attaching one counter to multiple rules, we are allocating counter per HW rule. We need this separation for two reasons: 1. "flow eswitch" counter affinity HW require the counter to be allocated on the device where the eswitch rule is set. 2. for some use-cases (multi-path routing) each HW flow relates to different neighbour, hence our neigh update logic must have a per-rule HW accountant in order to provide the proper feedback to the kernel. Signed-off-by: Shahar Klein Signed-off-by: Roi Dayan Reviewed-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 14 +++++++++++--- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 1 + 2 files changed, 12 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index eacccac05dda..779ca3a43bec 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -994,7 +994,7 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, } if (attr->action & MLX5_FLOW_CONTEXT_ACTION_COUNT) { - counter = mlx5_fc_create(esw->dev, true); + counter = mlx5_fc_create(attr->counter_dev, true); if (IS_ERR(counter)) { err = PTR_ERR(counter); goto err_create_counter; @@ -1023,7 +1023,7 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, return 0; err_add_rule: - mlx5_fc_destroy(esw->dev, counter); + mlx5_fc_destroy(attr->counter_dev, counter); err_create_counter: if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) mlx5e_detach_mod_hdr(priv, flow); @@ -1064,7 +1064,7 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv, mlx5e_detach_mod_hdr(priv, flow); if (attr->action & MLX5_FLOW_CONTEXT_ACTION_COUNT) - mlx5_fc_destroy(esw->dev, attr->counter); + mlx5_fc_destroy(attr->counter_dev, attr->counter); } void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv, @@ -2746,6 +2746,7 @@ __mlx5e_add_fdb_flow(struct mlx5e_priv *priv, struct mlx5e_tc_flow **__flow) { struct netlink_ext_ack *extack = f->common.extack; + struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; struct mlx5e_tc_flow_parse_attr *parse_attr; struct mlx5e_tc_flow *flow; int attr_size, err; @@ -2771,6 +2772,13 @@ __mlx5e_add_fdb_flow(struct mlx5e_priv *priv, flow->esw_attr->in_rep = in_rep; flow->esw_attr->in_mdev = in_mdev; + + if (MLX5_CAP_ESW(esw->dev, counter_eswitch_affinity) == + MLX5_COUNTER_SOURCE_ESWITCH) + flow->esw_attr->counter_dev = in_mdev; + else + flow->esw_attr->counter_dev = priv->mdev; + err = mlx5e_tc_add_fdb_flow(priv, parse_attr, flow, extack); if (err) goto err_free; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h index 4d048f7e703b..701abd794dab 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h @@ -293,6 +293,7 @@ enum { struct mlx5_esw_flow_attr { struct mlx5_eswitch_rep *in_rep; struct mlx5_core_dev *in_mdev; + struct mlx5_core_dev *counter_dev; int split_count; int out_count; -- cgit v1.2.3-55-g7522 From 491c37e49b48e7a18dc79e29f3ac13e0c3d0fb54 Mon Sep 17 00:00:00 2001 From: Rabie Loulou Date: Wed, 6 Jun 2018 15:49:27 +0300 Subject: net/mlx5e: In case of LAG, one switch parent id is used for all representors When the uplink representors are put into lag, set all the representors (VFs and uplinks) of the same NIC to return the same switchdev id. Currently, the route lookup code on the encapsulation offload path assumes that same switchdev id for the source and dest devices means that the dest is also mlx5 HW netdev. This doesn't hold anymore when we align the switchdev Id of the uplinks to be same, which in turn causes the bond/team to return that id to the caller. As such, enhance the relevant check to take into account the uplink lag case. Signed-off-by: Rabie Loulou Signed-off-by: Aviv Heller Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/en/tc_tun.c | 41 ++++++++++++++++------ drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 20 +++++++++-- 2 files changed, 48 insertions(+), 13 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') 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 c1515f013501..d5d161ab0dbc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c @@ -14,7 +14,8 @@ static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv, u8 *out_ttl) { struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; - struct mlx5e_rep_priv *uplink_rpriv; + struct net_device *uplink_dev, *uplink_upper; + bool dst_is_lag_dev; struct rtable *rt; struct neighbour *n = NULL; @@ -28,10 +29,20 @@ static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv, #else return -EOPNOTSUPP; #endif - uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH); - /* if the egress device isn't on the same HW e-switch, we use the uplink */ - if (!switchdev_port_same_parent_id(priv->netdev, rt->dst.dev)) - *out_dev = uplink_rpriv->netdev; + + uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH); + uplink_upper = netdev_master_upper_dev_get(uplink_dev); + dst_is_lag_dev = (uplink_upper && + netif_is_lag_master(uplink_upper) && + rt->dst.dev == uplink_upper && + mlx5_lag_is_active(priv->mdev)); + + /* if the egress device isn't on the same HW e-switch or + * it's a LAG device, use the uplink + */ + if (!switchdev_port_same_parent_id(priv->netdev, rt->dst.dev) || + dst_is_lag_dev) + *out_dev = uplink_dev; else *out_dev = rt->dst.dev; @@ -65,8 +76,9 @@ static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv, struct dst_entry *dst; #if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6) - struct mlx5e_rep_priv *uplink_rpriv; struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; + struct net_device *uplink_dev, *uplink_upper; + bool dst_is_lag_dev; int ret; ret = ipv6_stub->ipv6_dst_lookup(dev_net(mirred_dev), NULL, &dst, @@ -77,10 +89,19 @@ static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv, if (!(*out_ttl)) *out_ttl = ip6_dst_hoplimit(dst); - uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH); - /* if the egress device isn't on the same HW e-switch, we use the uplink */ - if (!switchdev_port_same_parent_id(priv->netdev, dst->dev)) - *out_dev = uplink_rpriv->netdev; + uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH); + uplink_upper = netdev_master_upper_dev_get(uplink_dev); + dst_is_lag_dev = (uplink_upper && + netif_is_lag_master(uplink_upper) && + dst->dev == uplink_upper && + mlx5_lag_is_active(priv->mdev)); + + /* if the egress device isn't on the same HW e-switch or + * it's a LAG device, use the uplink + */ + if (!switchdev_port_same_parent_id(priv->netdev, dst->dev) || + dst_is_lag_dev) + *out_dev = uplink_dev; else *out_dev = dst->dev; #else diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index e5230049612f..17f24127a3ba 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -297,17 +297,31 @@ static const struct ethtool_ops mlx5e_rep_ethtool_ops = { int mlx5e_attr_get(struct net_device *dev, struct switchdev_attr *attr) { struct mlx5e_priv *priv = netdev_priv(dev); - struct mlx5e_rep_priv *rpriv = priv->ppriv; - struct mlx5_eswitch_rep *rep = rpriv->rep; struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; + struct net_device *uplink_upper = NULL; + struct mlx5e_priv *uplink_priv = NULL; + struct net_device *uplink_dev; if (esw->mode == SRIOV_NONE) return -EOPNOTSUPP; + uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH); + if (uplink_dev) { + uplink_upper = netdev_master_upper_dev_get(uplink_dev); + uplink_priv = netdev_priv(uplink_dev); + } + switch (attr->id) { case SWITCHDEV_ATTR_ID_PORT_PARENT_ID: attr->u.ppid.id_len = ETH_ALEN; - ether_addr_copy(attr->u.ppid.id, rep->hw_id); + if (uplink_upper && mlx5_lag_is_active(uplink_priv->mdev)) { + ether_addr_copy(attr->u.ppid.id, uplink_upper->dev_addr); + } else { + struct mlx5e_rep_priv *rpriv = priv->ppriv; + struct mlx5_eswitch_rep *rep = rpriv->rep; + + ether_addr_copy(attr->u.ppid.id, rep->hw_id); + } break; default: return -EOPNOTSUPP; -- cgit v1.2.3-55-g7522 From 7ba58ba7ba5e19324b3879000ef341966b19727b Mon Sep 17 00:00:00 2001 From: Rabie Loulou Date: Wed, 6 Jun 2018 16:27:08 +0300 Subject: net/mlx5e: Offload TC e-switch rules with egress LAG device When parsing TC FDB actions, if the egress device is a bond/team net-device which enslaved the uplink representor of the e-switch, use the uplink representor as the destination in the HW rule. Signed-off-by: Rabie Loulou Signed-off-by: Aviv Heller Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 779ca3a43bec..cede77fd208f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -2574,6 +2574,15 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, if (switchdev_port_same_parent_id(priv->netdev, out_dev) || is_merged_eswitch_dev(priv, out_dev)) { + struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; + struct net_device *uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH); + struct net_device *uplink_upper = netdev_master_upper_dev_get(uplink_dev); + + if (uplink_upper && + netif_is_lag_master(uplink_upper) && + uplink_upper == out_dev) + out_dev = uplink_dev; + out_priv = netdev_priv(out_dev); rpriv = out_priv->ppriv; attr->dests[attr->out_count].rep = rpriv->rep; -- cgit v1.2.3-55-g7522 From 1418ddd96afdb097ab9a2fe09c94e820d596321a Mon Sep 17 00:00:00 2001 From: Aviv Heller Date: Tue, 2 Oct 2018 17:18:44 +0300 Subject: net/mlx5e: Duplicate offloaded TC eswitch rules under uplink LAG Under uplink LAG, packets that match a flow might arrive on both uplink ports and transmitted through both as part of supporting aggregation and high-availability. When the netdevs representing the uplinks are set into LAG (bonding, teaming), duplicate the TC flow offloading into each of the per-uplink e-switches. Duplication is not required if the source is the bond device, since in this case it is assumed that the bond and the uplink netdevs share the same TC block, and thus duplication will occur naturally by the stack. Note that under encapsulation scheme, both flows will use the same neighbour and hence both will contribute to the last-used feedback towards the stack. Signed-off-by: Aviv Heller Signed-off-by: Rabie Loulou Reviewed-by: Or Gerlitz Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index cede77fd208f..864f3b00d09d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -2710,7 +2710,16 @@ static struct rhashtable *get_tc_ht(struct mlx5e_priv *priv) static bool is_peer_flow_needed(struct mlx5e_tc_flow *flow) { - return false; + struct mlx5_esw_flow_attr *attr = flow->esw_attr; + bool is_rep_ingress = attr->in_rep->vport != FDB_UPLINK_VPORT && + flow->flags & MLX5E_TC_FLOW_INGRESS; + bool act_is_encap = !!(attr->action & + MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT); + bool esw_paired = mlx5_devcom_is_paired(attr->in_mdev->priv.devcom, + MLX5_DEVCOM_ESW_OFFLOADS); + + return esw_paired && mlx5_lag_is_active(attr->in_mdev) && + (is_rep_ingress || act_is_encap); } static int -- cgit v1.2.3-55-g7522 From 3b5ff59fd851d8e8c7c3ba08b01011baffa60cb6 Mon Sep 17 00:00:00 2001 From: Rabie Loulou Date: Thu, 26 Apr 2018 16:45:41 +0300 Subject: net/mlx5: Adjustments for the activate LAG logic to run under sriov When HW lag is set/unset, roce must not be enabled on the port, as such we wrap such changes with roce enable/disable either directly or through re-creation of IB device. Currently, lag and sriov are mutually exclusive, so by definition this code doesn't run under sriov. Towards changing this exclusion, we need to make sure that roce will not be enabled on the eswitch manager port under sriov since this is requirement of the switchdev mode. We are going strict here and avoiding this all together under sriov. Signed-off-by: Rabie Loulou Reviewed-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/lag.c | 33 +++++++++++++++++---------- 1 file changed, 21 insertions(+), 12 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c index 8c5c5e418d61..8127d907e1ef 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c @@ -34,6 +34,7 @@ #include #include #include "mlx5_core.h" +#include "eswitch.h" enum { MLX5_LAG_FLAG_BONDED = 1 << 0, @@ -257,13 +258,15 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) { struct mlx5_core_dev *dev0 = ldev->pf[0].dev; struct mlx5_core_dev *dev1 = ldev->pf[1].dev; + bool do_bond, sriov_enabled; struct lag_tracker tracker; int i; - bool do_bond; if (!dev0 || !dev1) return; + sriov_enabled = mlx5_sriov_is_enabled(dev0) || mlx5_sriov_is_enabled(dev1); + mutex_lock(&lag_mutex); tracker = ldev->tracker; mutex_unlock(&lag_mutex); @@ -271,26 +274,32 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) do_bond = tracker.is_bonded && ldev->allowed; if (do_bond && !mlx5_lag_is_bonded(ldev)) { - for (i = 0; i < MLX5_MAX_PORTS; i++) - mlx5_remove_dev_by_protocol(ldev->pf[i].dev, - MLX5_INTERFACE_PROTOCOL_IB); + if (!sriov_enabled) + for (i = 0; i < MLX5_MAX_PORTS; i++) + mlx5_remove_dev_by_protocol(ldev->pf[i].dev, + MLX5_INTERFACE_PROTOCOL_IB); mlx5_activate_lag(ldev, &tracker); - mlx5_add_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); - mlx5_nic_vport_enable_roce(dev1); + if (!sriov_enabled) { + mlx5_add_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); + mlx5_nic_vport_enable_roce(dev1); + } } else if (do_bond && mlx5_lag_is_bonded(ldev)) { mlx5_modify_lag(ldev, &tracker); } else if (!do_bond && mlx5_lag_is_bonded(ldev)) { - mlx5_remove_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); - mlx5_nic_vport_disable_roce(dev1); + if (!sriov_enabled) { + mlx5_remove_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); + mlx5_nic_vport_disable_roce(dev1); + } mlx5_deactivate_lag(ldev); - for (i = 0; i < MLX5_MAX_PORTS; i++) - if (ldev->pf[i].dev) - mlx5_add_dev_by_protocol(ldev->pf[i].dev, - MLX5_INTERFACE_PROTOCOL_IB); + if (!sriov_enabled) + for (i = 0; i < MLX5_MAX_PORTS; i++) + if (ldev->pf[i].dev) + mlx5_add_dev_by_protocol(ldev->pf[i].dev, + MLX5_INTERFACE_PROTOCOL_IB); } } -- cgit v1.2.3-55-g7522 From eff849b2c669ebf0df88762f082bcb252f8e883c Mon Sep 17 00:00:00 2001 From: Rabie Loulou Date: Wed, 6 Jun 2018 16:31:34 +0300 Subject: net/mlx5: Allow/disallow LAG according to pre-req only Remove the lag forbid/allow functions, change the lag prereq check to run in the do-bond logic, so every change in the prereq state will cause LAG to be disabled/enabled accordingly after the next do-bond run. Add lag update function, so every component which changes the prereq state and want the LAG to re-calc the conditions can call the update function. Signed-off-by: Rabie Loulou Signed-off-by: Aviv Heller Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 13 +++++ drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 4 ++ drivers/net/ethernet/mellanox/mlx5/core/lag.c | 62 ++++++---------------- .../net/ethernet/mellanox/mlx5/core/mlx5_core.h | 4 +- drivers/net/ethernet/mellanox/mlx5/core/sriov.c | 14 +---- 5 files changed, 36 insertions(+), 61 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index 4f9dd0d4a081..70d0d9aa6b5d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -1632,6 +1632,8 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode) esw_info(esw->dev, "E-Switch enable SRIOV: nvfs(%d) mode (%d)\n", nvfs, mode); esw->mode = mode; + mlx5_lag_update(esw->dev); + if (mode == SRIOV_LEGACY) { err = esw_create_legacy_fdb_table(esw); } else { @@ -1708,6 +1710,8 @@ void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw) old_mode = esw->mode; esw->mode = SRIOV_NONE; + mlx5_lag_update(esw->dev); + if (old_mode == SRIOV_OFFLOADS) mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB); } @@ -2225,3 +2229,12 @@ u8 mlx5_eswitch_mode(struct mlx5_eswitch *esw) return ESW_ALLOWED(esw) ? esw->mode : SRIOV_NONE; } EXPORT_SYMBOL_GPL(mlx5_eswitch_mode); + +bool mlx5_esw_lag_prereq(struct mlx5_core_dev *dev0, struct mlx5_core_dev *dev1) +{ + if (dev0->priv.eswitch->mode == SRIOV_NONE && + dev1->priv.eswitch->mode == SRIOV_NONE) + return true; + + return false; +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h index 701abd794dab..9c89eea9b2c3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h @@ -351,6 +351,9 @@ static inline bool mlx5_eswitch_vlan_actions_supported(struct mlx5_core_dev *dev MLX5_CAP_ESW_FLOWTABLE_FDB(dev, push_vlan_2); } +bool mlx5_esw_lag_prereq(struct mlx5_core_dev *dev0, + struct mlx5_core_dev *dev1); + #define MLX5_DEBUG_ESWITCH_MASK BIT(3) #define esw_info(dev, format, ...) \ @@ -367,6 +370,7 @@ static inline int mlx5_eswitch_init(struct mlx5_core_dev *dev) { return 0; } static inline void mlx5_eswitch_cleanup(struct mlx5_eswitch *esw) {} static inline int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode) { return 0; } static inline void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw) {} +static inline bool mlx5_esw_lag_prereq(struct mlx5_core_dev *dev0, struct mlx5_core_dev *dev1) { return true; } #define FDB_MAX_CHAIN 1 #define FDB_SLOW_PATH_CHAIN (FDB_MAX_CHAIN + 1) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c index 8127d907e1ef..3b1dcd2969d7 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c @@ -62,11 +62,6 @@ struct mlx5_lag { struct lag_tracker tracker; struct delayed_work bond_work; struct notifier_block nb; - - /* Admin state. Allow lag only if allowed is true - * even if network conditions for lag were met - */ - bool allowed; }; /* General purpose, use for short periods of time. @@ -254,6 +249,16 @@ static void mlx5_deactivate_lag(struct mlx5_lag *ldev) err); } +static bool mlx5_lag_check_prereq(struct mlx5_lag *ldev) +{ + if (ldev->pf[0].dev && + ldev->pf[1].dev && + mlx5_esw_lag_prereq(ldev->pf[0].dev, ldev->pf[1].dev)) + return true; + else + return false; +} + static void mlx5_do_bond(struct mlx5_lag *ldev) { struct mlx5_core_dev *dev0 = ldev->pf[0].dev; @@ -271,7 +276,7 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) tracker = ldev->tracker; mutex_unlock(&lag_mutex); - do_bond = tracker.is_bonded && ldev->allowed; + do_bond = tracker.is_bonded && mlx5_lag_check_prereq(ldev); if (do_bond && !mlx5_lag_is_bonded(ldev)) { if (!sriov_enabled) @@ -449,15 +454,6 @@ static int mlx5_lag_netdev_event(struct notifier_block *this, return NOTIFY_DONE; } -static bool mlx5_lag_check_prereq(struct mlx5_lag *ldev) -{ - if ((ldev->pf[0].dev && mlx5_sriov_is_enabled(ldev->pf[0].dev)) || - (ldev->pf[1].dev && mlx5_sriov_is_enabled(ldev->pf[1].dev))) - return false; - else - return true; -} - static struct mlx5_lag *mlx5_lag_dev_alloc(void) { struct mlx5_lag *ldev; @@ -467,7 +463,6 @@ static struct mlx5_lag *mlx5_lag_dev_alloc(void) return NULL; INIT_DELAYED_WORK(&ldev->bond_work, mlx5_do_bond_work); - ldev->allowed = mlx5_lag_check_prereq(ldev); return ldev; } @@ -492,7 +487,6 @@ static void mlx5_lag_dev_add_pf(struct mlx5_lag *ldev, ldev->tracker.netdev_state[fn].link_up = 0; ldev->tracker.netdev_state[fn].tx_enabled = 0; - ldev->allowed = mlx5_lag_check_prereq(ldev); dev->priv.lag = ldev; mutex_unlock(&lag_mutex); @@ -514,7 +508,6 @@ static void mlx5_lag_dev_remove_pf(struct mlx5_lag *ldev, memset(&ldev->pf[i], 0, sizeof(*ldev->pf)); dev->priv.lag = NULL; - ldev->allowed = mlx5_lag_check_prereq(ldev); mutex_unlock(&lag_mutex); } @@ -593,42 +586,19 @@ bool mlx5_lag_is_active(struct mlx5_core_dev *dev) } EXPORT_SYMBOL(mlx5_lag_is_active); -static int mlx5_lag_set_state(struct mlx5_core_dev *dev, bool allow) +void mlx5_lag_update(struct mlx5_core_dev *dev) { struct mlx5_lag *ldev; - int ret = 0; - bool lag_active; mlx5_dev_list_lock(); - ldev = mlx5_lag_dev_get(dev); - if (!ldev) { - ret = -ENODEV; - goto unlock; - } - lag_active = mlx5_lag_is_bonded(ldev); - if (!mlx5_lag_check_prereq(ldev) && allow) { - ret = -EINVAL; - goto unlock; - } - if (ldev->allowed == allow) + if (!ldev) goto unlock; - ldev->allowed = allow; - if ((lag_active && !allow) || allow) - mlx5_do_bond(ldev); -unlock: - mlx5_dev_list_unlock(); - return ret; -} -int mlx5_lag_forbid(struct mlx5_core_dev *dev) -{ - return mlx5_lag_set_state(dev, false); -} + mlx5_do_bond(ldev); -int mlx5_lag_allow(struct mlx5_core_dev *dev) -{ - return mlx5_lag_set_state(dev, true); +unlock: + mlx5_dev_list_unlock(); } struct net_device *mlx5_lag_get_roce_netdev(struct mlx5_core_dev *dev) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index fd3141a4b3f1..73bf46599ec6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -185,10 +185,8 @@ static inline int mlx5_lag_is_lacp_owner(struct mlx5_core_dev *dev) MLX5_CAP_GEN(dev, lag_master); } -int mlx5_lag_allow(struct mlx5_core_dev *dev); -int mlx5_lag_forbid(struct mlx5_core_dev *dev); - void mlx5_reload_interface(struct mlx5_core_dev *mdev, int protocol); +void mlx5_lag_update(struct mlx5_core_dev *dev); enum { MLX5_NIC_IFC_FULL = 0, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/sriov.c b/drivers/net/ethernet/mellanox/mlx5/core/sriov.c index a0674962f02c..6e178030d8fb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/sriov.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/sriov.c @@ -216,20 +216,10 @@ int mlx5_core_sriov_configure(struct pci_dev *pdev, int num_vfs) if (!mlx5_core_is_pf(dev)) return -EPERM; - if (num_vfs) { - int ret; - - ret = mlx5_lag_forbid(dev); - if (ret && (ret != -ENODEV)) - return ret; - } - - if (num_vfs) { + if (num_vfs) err = mlx5_sriov_enable(pdev, num_vfs); - } else { + else mlx5_sriov_disable(pdev); - mlx5_lag_allow(dev); - } return err ? err : num_vfs; } -- cgit v1.2.3-55-g7522 From 8aaca1976e2757e15edf43866a0549b321a4222f Mon Sep 17 00:00:00 2001 From: Rabie Loulou Date: Wed, 23 May 2018 14:19:07 +0300 Subject: net/mlx5: Allow co-enablement of uplink LAG and SRIOV Enable setting uplink LAG if sriov is enabled on both ports in switchdev mode. Once the sriov mode is changed from switchdev for any of the ports, the LAG instance is disabled. Signed-off-by: Rabie Loulou Signed-off-by: Aviv Heller Signed-off-by: Jianbo Liu Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index 70d0d9aa6b5d..ef2bec50423d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -2232,8 +2232,10 @@ EXPORT_SYMBOL_GPL(mlx5_eswitch_mode); bool mlx5_esw_lag_prereq(struct mlx5_core_dev *dev0, struct mlx5_core_dev *dev1) { - if (dev0->priv.eswitch->mode == SRIOV_NONE && - dev1->priv.eswitch->mode == SRIOV_NONE) + if ((dev0->priv.eswitch->mode == SRIOV_NONE && + dev1->priv.eswitch->mode == SRIOV_NONE) || + (dev0->priv.eswitch->mode == SRIOV_OFFLOADS && + dev1->priv.eswitch->mode == SRIOV_OFFLOADS)) return true; return false; -- cgit v1.2.3-55-g7522 From 292612d68c4ed3063cdc99512e865a819b8c9423 Mon Sep 17 00:00:00 2001 From: Aviv Heller Date: Thu, 23 Aug 2018 13:34:10 +0300 Subject: net/mlx5: Rename mlx5_lag_is_bonded() to __mlx5_lag_is_active() The new name better represents the function's aim, and sets a precedent for a '__' prefix for internal, non-locked versions of LAG APIs. Signed-off-by: Aviv Heller Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/lag.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c index 3b1dcd2969d7..db5ef7023371 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c @@ -161,7 +161,7 @@ static int mlx5_lag_dev_get_netdev_idx(struct mlx5_lag *ldev, return -1; } -static bool mlx5_lag_is_bonded(struct mlx5_lag *ldev) +static bool __mlx5_lag_is_active(struct mlx5_lag *ldev) { return !!(ldev->flags & MLX5_LAG_FLAG_BONDED); } @@ -278,7 +278,7 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) do_bond = tracker.is_bonded && mlx5_lag_check_prereq(ldev); - if (do_bond && !mlx5_lag_is_bonded(ldev)) { + if (do_bond && !__mlx5_lag_is_active(ldev)) { if (!sriov_enabled) for (i = 0; i < MLX5_MAX_PORTS; i++) mlx5_remove_dev_by_protocol(ldev->pf[i].dev, @@ -290,9 +290,9 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) mlx5_add_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); mlx5_nic_vport_enable_roce(dev1); } - } else if (do_bond && mlx5_lag_is_bonded(ldev)) { + } else if (do_bond && __mlx5_lag_is_active(ldev)) { mlx5_modify_lag(ldev, &tracker); - } else if (!do_bond && mlx5_lag_is_bonded(ldev)) { + } else if (!do_bond && __mlx5_lag_is_active(ldev)) { if (!sriov_enabled) { mlx5_remove_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); mlx5_nic_vport_disable_roce(dev1); @@ -555,7 +555,7 @@ void mlx5_lag_remove(struct mlx5_core_dev *dev) if (!ldev) return; - if (mlx5_lag_is_bonded(ldev)) + if (__mlx5_lag_is_active(ldev)) mlx5_deactivate_lag(ldev); mlx5_lag_dev_remove_pf(ldev, dev); @@ -579,7 +579,7 @@ bool mlx5_lag_is_active(struct mlx5_core_dev *dev) mutex_lock(&lag_mutex); ldev = mlx5_lag_dev_get(dev); - res = ldev && mlx5_lag_is_bonded(ldev); + res = ldev && __mlx5_lag_is_active(ldev); mutex_unlock(&lag_mutex); return res; @@ -609,7 +609,7 @@ struct net_device *mlx5_lag_get_roce_netdev(struct mlx5_core_dev *dev) mutex_lock(&lag_mutex); ldev = mlx5_lag_dev_get(dev); - if (!(ldev && mlx5_lag_is_bonded(ldev))) + if (!(ldev && __mlx5_lag_is_active(ldev))) goto unlock; if (ldev->tracker.tx_type == NETDEV_LAG_TX_TYPE_ACTIVEBACKUP) { @@ -638,7 +638,7 @@ bool mlx5_lag_intf_add(struct mlx5_interface *intf, struct mlx5_priv *priv) return true; ldev = mlx5_lag_dev_get(dev); - if (!ldev || !mlx5_lag_is_bonded(ldev) || ldev->pf[0].dev == dev) + if (!ldev || !__mlx5_lag_is_active(ldev) || ldev->pf[0].dev == dev) return true; /* If bonded, we do not add an IB device for PF1. */ @@ -665,7 +665,7 @@ int mlx5_lag_query_cong_counters(struct mlx5_core_dev *dev, mutex_lock(&lag_mutex); ldev = mlx5_lag_dev_get(dev); - if (ldev && mlx5_lag_is_bonded(ldev)) { + if (ldev && __mlx5_lag_is_active(ldev)) { num_ports = MLX5_MAX_PORTS; mdev[0] = ldev->pf[0].dev; mdev[1] = ldev->pf[1].dev; -- cgit v1.2.3-55-g7522 From 7c34ec19e10c0d13ca2f3435fb85d2dddccad917 Mon Sep 17 00:00:00 2001 From: Aviv Heller Date: Thu, 23 Aug 2018 13:47:53 +0300 Subject: net/mlx5: Make RoCE and SR-IOV LAG modes explicit With the introduction of SR-IOV LAG, checking whether LAG is active is no longer good enough, since RoCE and SR-IOV LAG each entails different behavior by both the core and infiniband drivers. This patch introduces facilities to discern LAG type, in addition to mlx5_lag_is_active(). These are implemented in such a way as to allow more complex mode combinations in the future. Signed-off-by: Aviv Heller Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- drivers/infiniband/hw/mlx5/main.c | 13 ++-- drivers/infiniband/hw/mlx5/mlx5_ib.h | 1 + drivers/infiniband/hw/mlx5/qp.c | 2 +- .../net/ethernet/mellanox/mlx5/core/en/tc_tun.c | 4 +- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/lag.c | 79 +++++++++++++++++----- include/linux/mlx5/driver.h | 2 + 8 files changed, 79 insertions(+), 26 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index d66457e6ffba..e85974ab06c0 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -445,7 +445,7 @@ static int mlx5_query_port_roce(struct ib_device *device, u8 port_num, if (!ndev) goto out; - if (mlx5_lag_is_active(dev->mdev)) { + if (dev->lag_active) { rcu_read_lock(); upper = netdev_master_upper_dev_get_rcu(ndev); if (upper) { @@ -1848,7 +1848,7 @@ static struct ib_ucontext *mlx5_ib_alloc_ucontext(struct ib_device *ibdev, context->lib_caps = req.lib_caps; print_lib_caps(dev, context->lib_caps); - if (mlx5_lag_is_active(dev->mdev)) { + if (dev->lag_active) { u8 port = mlx5_core_native_port_num(dev->mdev); atomic_set(&context->tx_port_affinity, @@ -4841,7 +4841,7 @@ static int mlx5_eth_lag_init(struct mlx5_ib_dev *dev) struct mlx5_flow_table *ft; int err; - if (!ns || !mlx5_lag_is_active(mdev)) + if (!ns || !mlx5_lag_is_roce(mdev)) return 0; err = mlx5_cmd_create_vport_lag(mdev); @@ -4855,6 +4855,7 @@ static int mlx5_eth_lag_init(struct mlx5_ib_dev *dev) } dev->flow_db->lag_demux_ft = ft; + dev->lag_active = true; return 0; err_destroy_vport_lag: @@ -4866,7 +4867,9 @@ static void mlx5_eth_lag_cleanup(struct mlx5_ib_dev *dev) { struct mlx5_core_dev *mdev = dev->mdev; - if (dev->flow_db->lag_demux_ft) { + if (dev->lag_active) { + dev->lag_active = false; + mlx5_destroy_flow_table(dev->flow_db->lag_demux_ft); dev->flow_db->lag_demux_ft = NULL; @@ -6173,7 +6176,7 @@ int mlx5_ib_stage_ib_reg_init(struct mlx5_ib_dev *dev) const char *name; rdma_set_device_sysfs_group(&dev->ib_dev, &mlx5_attr_group); - if (!mlx5_lag_is_active(dev->mdev)) + if (!mlx5_lag_is_roce(dev->mdev)) name = "mlx5_%d"; else name = "mlx5_bond_%d"; diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h index c89b3b44b22e..e507b6eb7c09 100644 --- a/drivers/infiniband/hw/mlx5/mlx5_ib.h +++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h @@ -936,6 +936,7 @@ struct mlx5_ib_dev { struct mlx5_ib_delay_drop delay_drop; const struct mlx5_ib_profile *profile; struct mlx5_eswitch_rep *rep; + int lag_active; struct mlx5_ib_lb_state lb; u8 umr_fence; diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c index 3747cc681b18..a0e9ff763d42 100644 --- a/drivers/infiniband/hw/mlx5/qp.c +++ b/drivers/infiniband/hw/mlx5/qp.c @@ -3258,7 +3258,7 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp, (ibqp->qp_type == IB_QPT_RAW_PACKET) || (ibqp->qp_type == IB_QPT_XRC_INI) || (ibqp->qp_type == IB_QPT_XRC_TGT)) { - if (mlx5_lag_is_active(dev->mdev)) { + if (dev->lag_active) { u8 p = mlx5_core_native_port_num(dev->mdev); tx_affinity = get_tx_affinity(dev, pd, base, p); context->flags |= cpu_to_be32(tx_affinity << 24); 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 d5d161ab0dbc..b92f8b3ff6b2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c @@ -35,7 +35,7 @@ static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv, dst_is_lag_dev = (uplink_upper && netif_is_lag_master(uplink_upper) && rt->dst.dev == uplink_upper && - mlx5_lag_is_active(priv->mdev)); + mlx5_lag_is_sriov(priv->mdev)); /* if the egress device isn't on the same HW e-switch or * it's a LAG device, use the uplink @@ -94,7 +94,7 @@ static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv, dst_is_lag_dev = (uplink_upper && netif_is_lag_master(uplink_upper) && dst->dev == uplink_upper && - mlx5_lag_is_active(priv->mdev)); + mlx5_lag_is_sriov(priv->mdev)); /* if the egress device isn't on the same HW e-switch or * it's a LAG device, use the uplink diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index 17f24127a3ba..e4a34c9ef700 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -314,7 +314,7 @@ int mlx5e_attr_get(struct net_device *dev, struct switchdev_attr *attr) switch (attr->id) { case SWITCHDEV_ATTR_ID_PORT_PARENT_ID: attr->u.ppid.id_len = ETH_ALEN; - if (uplink_upper && mlx5_lag_is_active(uplink_priv->mdev)) { + if (uplink_upper && mlx5_lag_is_sriov(uplink_priv->mdev)) { ether_addr_copy(attr->u.ppid.id, uplink_upper->dev_addr); } else { struct mlx5e_rep_priv *rpriv = priv->ppriv; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 864f3b00d09d..53ebb5a48018 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -2718,7 +2718,7 @@ static bool is_peer_flow_needed(struct mlx5e_tc_flow *flow) bool esw_paired = mlx5_devcom_is_paired(attr->in_mdev->priv.devcom, MLX5_DEVCOM_ESW_OFFLOADS); - return esw_paired && mlx5_lag_is_active(attr->in_mdev) && + return esw_paired && mlx5_lag_is_sriov(attr->in_mdev) && (is_rep_ingress || act_is_encap); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c index db5ef7023371..feb8230d3f86 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c @@ -37,9 +37,12 @@ #include "eswitch.h" enum { - MLX5_LAG_FLAG_BONDED = 1 << 0, + MLX5_LAG_FLAG_ROCE = 1 << 0, + MLX5_LAG_FLAG_SRIOV = 1 << 1, }; +#define MLX5_LAG_MODE_FLAGS (MLX5_LAG_FLAG_ROCE | MLX5_LAG_FLAG_SRIOV) + struct lag_func { struct mlx5_core_dev *dev; struct net_device *netdev; @@ -161,9 +164,19 @@ static int mlx5_lag_dev_get_netdev_idx(struct mlx5_lag *ldev, return -1; } +static bool __mlx5_lag_is_roce(struct mlx5_lag *ldev) +{ + return !!(ldev->flags & MLX5_LAG_FLAG_ROCE); +} + +static bool __mlx5_lag_is_sriov(struct mlx5_lag *ldev) +{ + return !!(ldev->flags & MLX5_LAG_FLAG_SRIOV); +} + static bool __mlx5_lag_is_active(struct mlx5_lag *ldev) { - return !!(ldev->flags & MLX5_LAG_FLAG_BONDED); + return !!(ldev->flags & MLX5_LAG_MODE_FLAGS); } static void mlx5_infer_tx_affinity_mapping(struct lag_tracker *tracker, @@ -229,9 +242,10 @@ static int mlx5_create_lag(struct mlx5_lag *ldev, } static void mlx5_activate_lag(struct mlx5_lag *ldev, - struct lag_tracker *tracker) + struct lag_tracker *tracker, + u8 flags) { - ldev->flags |= MLX5_LAG_FLAG_BONDED; + ldev->flags |= flags; mlx5_create_lag(ldev, tracker); } @@ -240,7 +254,7 @@ static void mlx5_deactivate_lag(struct mlx5_lag *ldev) struct mlx5_core_dev *dev0 = ldev->pf[0].dev; int err; - ldev->flags &= ~MLX5_LAG_FLAG_BONDED; + ldev->flags &= ~MLX5_LAG_MODE_FLAGS; err = mlx5_cmd_destroy_lag(dev0); if (err) @@ -263,15 +277,13 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) { struct mlx5_core_dev *dev0 = ldev->pf[0].dev; struct mlx5_core_dev *dev1 = ldev->pf[1].dev; - bool do_bond, sriov_enabled; struct lag_tracker tracker; + bool do_bond, roce_lag; int i; if (!dev0 || !dev1) return; - sriov_enabled = mlx5_sriov_is_enabled(dev0) || mlx5_sriov_is_enabled(dev1); - mutex_lock(&lag_mutex); tracker = ldev->tracker; mutex_unlock(&lag_mutex); @@ -279,28 +291,35 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) do_bond = tracker.is_bonded && mlx5_lag_check_prereq(ldev); if (do_bond && !__mlx5_lag_is_active(ldev)) { - if (!sriov_enabled) + roce_lag = !mlx5_sriov_is_enabled(dev0) && + !mlx5_sriov_is_enabled(dev1); + + if (roce_lag) for (i = 0; i < MLX5_MAX_PORTS; i++) mlx5_remove_dev_by_protocol(ldev->pf[i].dev, MLX5_INTERFACE_PROTOCOL_IB); - mlx5_activate_lag(ldev, &tracker); + mlx5_activate_lag(ldev, &tracker, + roce_lag ? MLX5_LAG_FLAG_ROCE : + MLX5_LAG_FLAG_SRIOV); - if (!sriov_enabled) { + if (roce_lag) { mlx5_add_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); mlx5_nic_vport_enable_roce(dev1); } } else if (do_bond && __mlx5_lag_is_active(ldev)) { mlx5_modify_lag(ldev, &tracker); } else if (!do_bond && __mlx5_lag_is_active(ldev)) { - if (!sriov_enabled) { + roce_lag = __mlx5_lag_is_roce(ldev); + + if (roce_lag) { mlx5_remove_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); mlx5_nic_vport_disable_roce(dev1); } mlx5_deactivate_lag(ldev); - if (!sriov_enabled) + if (roce_lag) for (i = 0; i < MLX5_MAX_PORTS; i++) if (ldev->pf[i].dev) mlx5_add_dev_by_protocol(ldev->pf[i].dev, @@ -572,6 +591,20 @@ void mlx5_lag_remove(struct mlx5_core_dev *dev) } } +bool mlx5_lag_is_roce(struct mlx5_core_dev *dev) +{ + struct mlx5_lag *ldev; + bool res; + + mutex_lock(&lag_mutex); + ldev = mlx5_lag_dev_get(dev); + res = ldev && __mlx5_lag_is_roce(ldev); + mutex_unlock(&lag_mutex); + + return res; +} +EXPORT_SYMBOL(mlx5_lag_is_roce); + bool mlx5_lag_is_active(struct mlx5_core_dev *dev) { struct mlx5_lag *ldev; @@ -586,6 +619,20 @@ bool mlx5_lag_is_active(struct mlx5_core_dev *dev) } EXPORT_SYMBOL(mlx5_lag_is_active); +bool mlx5_lag_is_sriov(struct mlx5_core_dev *dev) +{ + struct mlx5_lag *ldev; + bool res; + + mutex_lock(&lag_mutex); + ldev = mlx5_lag_dev_get(dev); + res = ldev && __mlx5_lag_is_sriov(ldev); + mutex_unlock(&lag_mutex); + + return res; +} +EXPORT_SYMBOL(mlx5_lag_is_sriov); + void mlx5_lag_update(struct mlx5_core_dev *dev) { struct mlx5_lag *ldev; @@ -609,7 +656,7 @@ struct net_device *mlx5_lag_get_roce_netdev(struct mlx5_core_dev *dev) mutex_lock(&lag_mutex); ldev = mlx5_lag_dev_get(dev); - if (!(ldev && __mlx5_lag_is_active(ldev))) + if (!(ldev && __mlx5_lag_is_roce(ldev))) goto unlock; if (ldev->tracker.tx_type == NETDEV_LAG_TX_TYPE_ACTIVEBACKUP) { @@ -638,7 +685,7 @@ bool mlx5_lag_intf_add(struct mlx5_interface *intf, struct mlx5_priv *priv) return true; ldev = mlx5_lag_dev_get(dev); - if (!ldev || !__mlx5_lag_is_active(ldev) || ldev->pf[0].dev == dev) + if (!ldev || !__mlx5_lag_is_roce(ldev) || ldev->pf[0].dev == dev) return true; /* If bonded, we do not add an IB device for PF1. */ @@ -665,7 +712,7 @@ int mlx5_lag_query_cong_counters(struct mlx5_core_dev *dev, mutex_lock(&lag_mutex); ldev = mlx5_lag_dev_get(dev); - if (ldev && __mlx5_lag_is_active(ldev)) { + if (ldev && __mlx5_lag_is_roce(ldev)) { num_ports = MLX5_MAX_PORTS; mdev[0] = ldev->pf[0].dev; mdev[1] = ldev->pf[1].dev; diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index cd7af5d0311b..4d16ba04790e 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -1019,6 +1019,8 @@ int mlx5_core_query_vendor_id(struct mlx5_core_dev *mdev, u32 *vendor_id); int mlx5_cmd_create_vport_lag(struct mlx5_core_dev *dev); int mlx5_cmd_destroy_vport_lag(struct mlx5_core_dev *dev); +bool mlx5_lag_is_roce(struct mlx5_core_dev *dev); +bool mlx5_lag_is_sriov(struct mlx5_core_dev *dev); bool mlx5_lag_is_active(struct mlx5_core_dev *dev); struct net_device *mlx5_lag_get_roce_netdev(struct mlx5_core_dev *dev); int mlx5_lag_query_cong_counters(struct mlx5_core_dev *dev, -- cgit v1.2.3-55-g7522 From 9582466640432c86fb060a4fa4d9d3c48e708ca6 Mon Sep 17 00:00:00 2001 From: Aviv Heller Date: Mon, 27 Aug 2018 15:39:40 +0300 Subject: net/mlx5: Handle LAG FW commands failure gracefully When create_lag or destroy_lag FW commands fail, display an appropriate error message, and try to recover, if possible. Signed-off-by: Aviv Heller Reviewed-by: Roi Dayan Reviewed-by: Yevgeny Kliteynik Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/lag.c | 91 ++++++++++++++++++++------- 1 file changed, 70 insertions(+), 21 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c index feb8230d3f86..5187dc7a72a3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c @@ -241,26 +241,52 @@ static int mlx5_create_lag(struct mlx5_lag *ldev, return err; } -static void mlx5_activate_lag(struct mlx5_lag *ldev, - struct lag_tracker *tracker, - u8 flags) +static int mlx5_activate_lag(struct mlx5_lag *ldev, + struct lag_tracker *tracker, + u8 flags) { + bool roce_lag = !!(flags & MLX5_LAG_FLAG_ROCE); + struct mlx5_core_dev *dev0 = ldev->pf[0].dev; + int err; + + err = mlx5_create_lag(ldev, tracker); + if (err) { + if (roce_lag) { + mlx5_core_err(dev0, + "Failed to activate RoCE LAG\n"); + } else { + mlx5_core_err(dev0, + "Failed to activate VF LAG\n" + "Make sure all VFs are unbound prior to VF LAG activation or deactivation\n"); + } + return err; + } + ldev->flags |= flags; - mlx5_create_lag(ldev, tracker); + return 0; } -static void mlx5_deactivate_lag(struct mlx5_lag *ldev) +static int mlx5_deactivate_lag(struct mlx5_lag *ldev) { struct mlx5_core_dev *dev0 = ldev->pf[0].dev; + bool roce_lag = __mlx5_lag_is_roce(ldev); int err; ldev->flags &= ~MLX5_LAG_MODE_FLAGS; err = mlx5_cmd_destroy_lag(dev0); - if (err) - mlx5_core_err(dev0, - "Failed to destroy LAG (%d)\n", - err); + if (err) { + if (roce_lag) { + mlx5_core_err(dev0, + "Failed to deactivate RoCE LAG; driver restart required\n"); + } else { + mlx5_core_err(dev0, + "Failed to deactivate VF LAG; driver restart required\n" + "Make sure all VFs are unbound prior to VF LAG activation or deactivation\n"); + } + } + + return err; } static bool mlx5_lag_check_prereq(struct mlx5_lag *ldev) @@ -273,13 +299,33 @@ static bool mlx5_lag_check_prereq(struct mlx5_lag *ldev) return false; } +static void mlx5_lag_add_ib_devices(struct mlx5_lag *ldev) +{ + int i; + + for (i = 0; i < MLX5_MAX_PORTS; i++) + if (ldev->pf[i].dev) + mlx5_add_dev_by_protocol(ldev->pf[i].dev, + MLX5_INTERFACE_PROTOCOL_IB); +} + +static void mlx5_lag_remove_ib_devices(struct mlx5_lag *ldev) +{ + int i; + + for (i = 0; i < MLX5_MAX_PORTS; i++) + if (ldev->pf[i].dev) + mlx5_remove_dev_by_protocol(ldev->pf[i].dev, + MLX5_INTERFACE_PROTOCOL_IB); +} + static void mlx5_do_bond(struct mlx5_lag *ldev) { struct mlx5_core_dev *dev0 = ldev->pf[0].dev; struct mlx5_core_dev *dev1 = ldev->pf[1].dev; struct lag_tracker tracker; bool do_bond, roce_lag; - int i; + int err; if (!dev0 || !dev1) return; @@ -295,13 +341,17 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) !mlx5_sriov_is_enabled(dev1); if (roce_lag) - for (i = 0; i < MLX5_MAX_PORTS; i++) - mlx5_remove_dev_by_protocol(ldev->pf[i].dev, - MLX5_INTERFACE_PROTOCOL_IB); + mlx5_lag_remove_ib_devices(ldev); + + err = mlx5_activate_lag(ldev, &tracker, + roce_lag ? MLX5_LAG_FLAG_ROCE : + MLX5_LAG_FLAG_SRIOV); + if (err) { + if (roce_lag) + mlx5_lag_add_ib_devices(ldev); - mlx5_activate_lag(ldev, &tracker, - roce_lag ? MLX5_LAG_FLAG_ROCE : - MLX5_LAG_FLAG_SRIOV); + return; + } if (roce_lag) { mlx5_add_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); @@ -317,13 +367,12 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) mlx5_nic_vport_disable_roce(dev1); } - mlx5_deactivate_lag(ldev); + err = mlx5_deactivate_lag(ldev); + if (err) + return; if (roce_lag) - for (i = 0; i < MLX5_MAX_PORTS; i++) - if (ldev->pf[i].dev) - mlx5_add_dev_by_protocol(ldev->pf[i].dev, - MLX5_INTERFACE_PROTOCOL_IB); + mlx5_lag_add_ib_devices(ldev); } } -- cgit v1.2.3-55-g7522 From 915fe1a0d9cbf6d7657b388833dbcae13d22a567 Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Tue, 13 Nov 2018 11:09:30 +0200 Subject: net/mlx5: E-Switch, Remove redundant reloading of the IB interface The reload of the IB interface done on the offloads stop call is redundant b/c we do that on mlx5_eswitch_disable_sriov(), remove it. Signed-off-by: Or Gerlitz Reviewed-by: Mark Bloch Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 76cb57202474..53065b6ae593 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -1450,9 +1450,6 @@ static int esw_offloads_stop(struct mlx5_eswitch *esw, } } - /* enable back PF RoCE */ - mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB); - return err; } -- cgit v1.2.3-55-g7522 From 025380b20dc2a3b92a36394799157a63a2fe5814 Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Wed, 28 Nov 2018 18:56:41 +0200 Subject: net/mlx5e: Use single argument for the esw representor build params helper This is prep step towards adding dedicated uplink representor. The patch doesn't change any functionality. Signed-off-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index e4a34c9ef700..996a03de9171 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1205,17 +1205,19 @@ static const struct net_device_ops mlx5e_netdev_ops_rep = { .ndo_change_mtu = mlx5e_change_rep_mtu, }; -static void mlx5e_build_rep_params(struct mlx5_core_dev *mdev, - struct mlx5e_params *params, - struct mlx5e_rss_params *rss_params, - u16 mtu) +static void mlx5e_build_rep_params(struct net_device *netdev) { + struct mlx5e_priv *priv = netdev_priv(netdev); + struct mlx5_core_dev *mdev = priv->mdev; + struct mlx5e_params *params; + u8 cq_period_mode = MLX5_CAP_GEN(mdev, cq_period_start_from_cqe) ? MLX5_CQ_PERIOD_MODE_START_FROM_CQE : MLX5_CQ_PERIOD_MODE_START_FROM_EQE; + params = &priv->channels.params; params->hard_mtu = MLX5E_ETH_HARD_MTU; - params->sw_mtu = mtu; + params->sw_mtu = netdev->mtu; params->log_sq_size = MLX5E_REP_PARAMS_LOG_SQ_SIZE; /* RQ */ @@ -1230,7 +1232,7 @@ static void mlx5e_build_rep_params(struct mlx5_core_dev *mdev, mlx5_query_min_inline(mdev, ¶ms->tx_min_inline_mode); /* RSS */ - mlx5e_build_rss_params(rss_params, params->num_channels); + mlx5e_build_rss_params(&priv->rss_params, params->num_channels); } static void mlx5e_build_rep_netdev(struct net_device *netdev) @@ -1283,8 +1285,7 @@ static int mlx5e_init_rep(struct mlx5_core_dev *mdev, priv->channels.params.num_channels = mlx5e_get_netdev_max_channels(netdev); - mlx5e_build_rep_params(mdev, &priv->channels.params, - &priv->rss_params, netdev->mtu); + mlx5e_build_rep_params(netdev); mlx5e_build_rep_netdev(netdev); mlx5e_timestamp_init(priv); -- cgit v1.2.3-55-g7522 From d9ee0491c2ffbb5bc8bc418dd857a045d0d0130f Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Tue, 13 Feb 2018 15:48:30 +0200 Subject: net/mlx5e: Use dedicated uplink vport netdev representor Currently, when running in sriov switchdev mode, we are using the PF netdevice as the uplink representor, this is problematic from few aspects: - will break when the PF isn't eswitch manager (e.g smart NIC env) - misalignment with other NIC switchdev drivers - makes us have and maintain special code, hurts the driver quality/robustness - which in turn opens the door for future bugs As of each and all of the above, we move to have a dedicated netdev representor for the uplink vport in a similar manner done for for the VF vports. This includes the following: 1. have an uplink rep netdev as we have for VF reps 2. all reps use same load/unload functions 3. HW stats for uplink based on physical port counters and not vport counters 4. link state for the uplink managed through PAOS and not vport state 5. the uplink rep has sysfs link to the PF PCI function && uses the PF MAC address Signed-off-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 3 +- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 31 +-- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 236 ++++++++++++++------- drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | 13 +- drivers/net/ethernet/mellanox/mlx5/core/en_stats.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/en_stats.h | 1 + drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 39 ++-- drivers/net/ethernet/mellanox/mlx5/core/en_tc.h | 8 +- 8 files changed, 212 insertions(+), 121 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 1c4e033b7509..31deafecf679 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -775,6 +775,7 @@ mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe, struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt); void mlx5e_update_stats(struct mlx5e_priv *priv); +void mlx5e_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats); void mlx5e_init_l2_addr(struct mlx5e_priv *priv); int mlx5e_self_test_num(struct mlx5e_priv *priv); @@ -939,7 +940,6 @@ int mlx5e_create_tis(struct mlx5_core_dev *mdev, int tc, void mlx5e_destroy_tis(struct mlx5_core_dev *mdev, u32 tisn); int mlx5e_create_tises(struct mlx5e_priv *priv); -void mlx5e_cleanup_nic_tx(struct mlx5e_priv *priv); int mlx5e_close(struct net_device *netdev); int mlx5e_open(struct net_device *netdev); void mlx5e_update_ndo_stats(struct mlx5e_priv *priv); @@ -948,6 +948,7 @@ void mlx5e_queue_update_stats(struct mlx5e_priv *priv); int mlx5e_bits_invert(unsigned long a, int size); typedef int (*change_hw_mtu_cb)(struct mlx5e_priv *priv); +int mlx5e_set_dev_port_mtu(struct mlx5e_priv *priv); int mlx5e_change_mtu(struct net_device *netdev, int new_mtu, change_hw_mtu_cb set_mtu_cb); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 59745851b5ba..cdd8b5efd96a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -2821,7 +2821,7 @@ static void mlx5e_query_mtu(struct mlx5_core_dev *mdev, *mtu = MLX5E_HW2SW_MTU(params, hw_mtu); } -static int mlx5e_set_dev_port_mtu(struct mlx5e_priv *priv) +int mlx5e_set_dev_port_mtu(struct mlx5e_priv *priv) { struct mlx5e_params *params = &priv->channels.params; struct net_device *netdev = priv->netdev; @@ -2901,7 +2901,7 @@ void mlx5e_activate_priv_channels(struct mlx5e_priv *priv) mlx5e_activate_channels(&priv->channels); netif_tx_start_all_queues(priv->netdev); - if (MLX5_ESWITCH_MANAGER(priv->mdev)) + if (mlx5e_is_vport_rep(priv)) mlx5e_add_sqs_fwd_rules(priv); mlx5e_wait_channels_min_rx_wqes(&priv->channels); @@ -2912,7 +2912,7 @@ void mlx5e_deactivate_priv_channels(struct mlx5e_priv *priv) { mlx5e_redirect_rqts_to_drop(priv); - if (MLX5_ESWITCH_MANAGER(priv->mdev)) + if (mlx5e_is_vport_rep(priv)) mlx5e_remove_sqs_fwd_rules(priv); /* FIXME: This is a W/A only for tx timeout watch dog false alarm when @@ -3164,7 +3164,7 @@ err_close_tises: return err; } -void mlx5e_cleanup_nic_tx(struct mlx5e_priv *priv) +static void mlx5e_cleanup_nic_tx(struct mlx5e_priv *priv) { int tc; @@ -3409,7 +3409,8 @@ static int mlx5e_setup_tc_block_cb(enum tc_setup_type type, void *type_data, switch (type) { case TC_SETUP_CLSFLOWER: - return mlx5e_setup_tc_cls_flower(priv, type_data, MLX5E_TC_INGRESS); + return mlx5e_setup_tc_cls_flower(priv, type_data, MLX5E_TC_INGRESS | + MLX5E_TC_NIC_OFFLOAD); default: return -EOPNOTSUPP; } @@ -3452,7 +3453,7 @@ static int mlx5e_setup_tc(struct net_device *dev, enum tc_setup_type type, } } -static void +void mlx5e_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats) { struct mlx5e_priv *priv = netdev_priv(dev); @@ -3596,7 +3597,7 @@ static int set_feature_tc_num_filters(struct net_device *netdev, bool enable) { struct mlx5e_priv *priv = netdev_priv(netdev); - if (!enable && mlx5e_tc_num_filters(priv)) { + if (!enable && mlx5e_tc_num_filters(priv, MLX5E_TC_NIC_OFFLOAD)) { netdev_err(netdev, "Active offloaded tc filters, can't turn hw_tc_offload off\n"); return -EINVAL; @@ -4607,12 +4608,6 @@ static void mlx5e_set_netdev_dev_addr(struct net_device *netdev) } } -#if IS_ENABLED(CONFIG_MLX5_ESWITCH) -static const struct switchdev_ops mlx5e_switchdev_ops = { - .switchdev_port_attr_get = mlx5e_attr_get, -}; -#endif - static void mlx5e_build_nic_netdev(struct net_device *netdev) { struct mlx5e_priv *priv = netdev_priv(netdev); @@ -4722,12 +4717,6 @@ static void mlx5e_build_nic_netdev(struct net_device *netdev) netdev->priv_flags |= IFF_UNICAST_FLT; mlx5e_set_netdev_dev_addr(netdev); - -#if IS_ENABLED(CONFIG_MLX5_ESWITCH) - if (MLX5_ESWITCH_MANAGER(mdev)) - netdev->switchdev_ops = &mlx5e_switchdev_ops; -#endif - mlx5e_ipsec_build_netdev(priv); mlx5e_tls_build_netdev(priv); } @@ -4908,7 +4897,7 @@ static void mlx5e_nic_enable(struct mlx5e_priv *priv) mlx5e_monitor_counter_init(priv); if (MLX5_ESWITCH_MANAGER(priv->mdev)) - mlx5e_register_vport_reps(priv); + mlx5e_rep_register_vport_reps(priv); if (netdev->reg_state != NETREG_REGISTERED) return; @@ -4943,7 +4932,7 @@ static void mlx5e_nic_disable(struct mlx5e_priv *priv) queue_work(priv->wq, &priv->set_rx_mode_work); if (MLX5_ESWITCH_MANAGER(priv->mdev)) - mlx5e_unregister_vport_reps(priv); + mlx5e_rep_unregister_vport_reps(priv); if (mlx5e_monitor_counter_supported(priv)) mlx5e_monitor_counter_cleanup(priv); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index 996a03de9171..b48ba3234ac5 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -108,7 +108,7 @@ static void mlx5e_rep_get_strings(struct net_device *dev, } } -static void mlx5e_rep_update_hw_counters(struct mlx5e_priv *priv) +static void mlx5e_vf_rep_update_hw_counters(struct mlx5e_priv *priv) { struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; struct mlx5e_rep_priv *rpriv = priv->ppriv; @@ -131,6 +131,32 @@ static void mlx5e_rep_update_hw_counters(struct mlx5e_priv *priv) vport_stats->tx_bytes = vf_stats.rx_bytes; } +static void mlx5e_uplink_rep_update_hw_counters(struct mlx5e_priv *priv) +{ + struct mlx5e_pport_stats *pstats = &priv->stats.pport; + struct rtnl_link_stats64 *vport_stats; + + mlx5e_grp_802_3_update_stats(priv); + + vport_stats = &priv->stats.vf_vport; + + vport_stats->rx_packets = PPORT_802_3_GET(pstats, a_frames_received_ok); + vport_stats->rx_bytes = PPORT_802_3_GET(pstats, a_octets_received_ok); + vport_stats->tx_packets = PPORT_802_3_GET(pstats, a_frames_transmitted_ok); + vport_stats->tx_bytes = PPORT_802_3_GET(pstats, a_octets_transmitted_ok); +} + +static void mlx5e_rep_update_hw_counters(struct mlx5e_priv *priv) +{ + struct mlx5e_rep_priv *rpriv = priv->ppriv; + struct mlx5_eswitch_rep *rep = rpriv->rep; + + if (rep->vport == FDB_UPLINK_VPORT) + mlx5e_uplink_rep_update_hw_counters(priv); + else + mlx5e_vf_rep_update_hw_counters(priv); +} + static void mlx5e_rep_update_sw_counters(struct mlx5e_priv *priv) { struct mlx5e_sw_stats *s = &priv->stats.sw; @@ -294,7 +320,7 @@ static const struct ethtool_ops mlx5e_rep_ethtool_ops = { .get_rxfh_indir_size = mlx5e_rep_get_rxfh_indir_size, }; -int mlx5e_attr_get(struct net_device *dev, struct switchdev_attr *attr) +static int mlx5e_attr_get(struct net_device *dev, struct switchdev_attr *attr) { struct mlx5e_priv *priv = netdev_priv(dev); struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; @@ -576,21 +602,19 @@ mlx5e_rep_indr_offload(struct net_device *netdev, struct tc_cls_flower_offload *flower, struct mlx5e_rep_indr_block_priv *indr_priv) { - int err = 0; struct mlx5e_priv *priv = netdev_priv(indr_priv->rpriv->netdev); + int flags = MLX5E_TC_EGRESS | MLX5E_TC_ESW_OFFLOAD; + int err = 0; switch (flower->command) { case TC_CLSFLOWER_REPLACE: - err = mlx5e_configure_flower(netdev, priv, - flower, MLX5E_TC_EGRESS); + err = mlx5e_configure_flower(netdev, priv, flower, flags); break; case TC_CLSFLOWER_DESTROY: - err = mlx5e_delete_flower(netdev, priv, - flower, MLX5E_TC_EGRESS); + err = mlx5e_delete_flower(netdev, priv, flower, flags); break; case TC_CLSFLOWER_STATS: - err = mlx5e_stats_flower(netdev, priv, - flower, MLX5E_TC_EGRESS); + err = mlx5e_stats_flower(netdev, priv, flower, flags); break; default: err = -EOPNOTSUPP; @@ -983,7 +1007,7 @@ void mlx5e_rep_encap_entry_detach(struct mlx5e_priv *priv, mlx5e_rep_neigh_entry_destroy(priv, nhe); } -static int mlx5e_rep_open(struct net_device *dev) +static int mlx5e_vf_rep_open(struct net_device *dev) { struct mlx5e_priv *priv = netdev_priv(dev); struct mlx5e_rep_priv *rpriv = priv->ppriv; @@ -1005,7 +1029,7 @@ unlock: return err; } -static int mlx5e_rep_close(struct net_device *dev) +static int mlx5e_vf_rep_close(struct net_device *dev) { struct mlx5e_priv *priv = netdev_priv(dev); struct mlx5e_rep_priv *rpriv = priv->ppriv; @@ -1062,7 +1086,8 @@ static int mlx5e_rep_setup_tc_cb(enum tc_setup_type type, void *type_data, switch (type) { case TC_SETUP_CLSFLOWER: - return mlx5e_rep_setup_tc_cls_flower(priv, type_data, MLX5E_TC_INGRESS); + return mlx5e_rep_setup_tc_cls_flower(priv, type_data, MLX5E_TC_INGRESS | + MLX5E_TC_ESW_OFFLOAD); default: return -EOPNOTSUPP; } @@ -1101,19 +1126,17 @@ static int mlx5e_rep_setup_tc(struct net_device *dev, enum tc_setup_type type, bool mlx5e_is_uplink_rep(struct mlx5e_priv *priv) { - struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; struct mlx5e_rep_priv *rpriv = priv->ppriv; struct mlx5_eswitch_rep *rep; if (!MLX5_ESWITCH_MANAGER(priv->mdev)) return false; - rep = rpriv->rep; - if (esw->mode == SRIOV_OFFLOADS && - rep && rep->vport == FDB_UPLINK_VPORT) - return true; + if (!rpriv) /* non vport rep mlx5e instances don't use this field */ + return false; - return false; + rep = rpriv->rep; + return (rep->vport == FDB_UPLINK_VPORT); } static bool mlx5e_is_vf_vport_rep(struct mlx5e_priv *priv) @@ -1133,11 +1156,8 @@ static bool mlx5e_is_vf_vport_rep(struct mlx5e_priv *priv) bool mlx5e_has_offload_stats(const struct net_device *dev, int attr_id) { - struct mlx5e_priv *priv = netdev_priv(dev); - switch (attr_id) { case IFLA_OFFLOAD_XSTATS_CPU_HIT: - if (mlx5e_is_vf_vport_rep(priv) || mlx5e_is_uplink_rep(priv)) return true; } @@ -1175,7 +1195,7 @@ int mlx5e_get_offload_stats(int attr_id, const struct net_device *dev, } static void -mlx5e_rep_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats) +mlx5e_vf_rep_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats) { struct mlx5e_priv *priv = netdev_priv(dev); @@ -1184,30 +1204,59 @@ mlx5e_rep_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats) memcpy(stats, &priv->stats.vf_vport, sizeof(*stats)); } +static int mlx5e_vf_rep_change_mtu(struct net_device *netdev, int new_mtu) +{ + return mlx5e_change_mtu(netdev, new_mtu, NULL); +} + +static int mlx5e_uplink_rep_open(struct net_device *dev) +{ + int err; + + err = mlx5e_open(dev); + if (!err) + netif_carrier_on(dev); + return err; +} + +static int mlx5e_uplink_rep_change_mtu(struct net_device *netdev, int new_mtu) +{ + return mlx5e_change_mtu(netdev, new_mtu, mlx5e_set_dev_port_mtu); +} + static const struct switchdev_ops mlx5e_rep_switchdev_ops = { .switchdev_port_attr_get = mlx5e_attr_get, }; -static int mlx5e_change_rep_mtu(struct net_device *netdev, int new_mtu) -{ - return mlx5e_change_mtu(netdev, new_mtu, NULL); -} +static const struct net_device_ops mlx5e_netdev_ops_vf_rep = { + .ndo_open = mlx5e_vf_rep_open, + .ndo_stop = mlx5e_vf_rep_close, + .ndo_start_xmit = mlx5e_xmit, + .ndo_get_phys_port_name = mlx5e_rep_get_phys_port_name, + .ndo_setup_tc = mlx5e_rep_setup_tc, + .ndo_get_stats64 = mlx5e_vf_rep_get_stats, + .ndo_has_offload_stats = mlx5e_has_offload_stats, + .ndo_get_offload_stats = mlx5e_get_offload_stats, + .ndo_change_mtu = mlx5e_vf_rep_change_mtu, +}; -static const struct net_device_ops mlx5e_netdev_ops_rep = { - .ndo_open = mlx5e_rep_open, - .ndo_stop = mlx5e_rep_close, +static const struct net_device_ops mlx5e_netdev_ops_uplink_rep = { + .ndo_open = mlx5e_uplink_rep_open, + .ndo_stop = mlx5e_close, .ndo_start_xmit = mlx5e_xmit, .ndo_get_phys_port_name = mlx5e_rep_get_phys_port_name, .ndo_setup_tc = mlx5e_rep_setup_tc, - .ndo_get_stats64 = mlx5e_rep_get_stats, + .ndo_get_stats64 = mlx5e_get_stats, .ndo_has_offload_stats = mlx5e_has_offload_stats, .ndo_get_offload_stats = mlx5e_get_offload_stats, - .ndo_change_mtu = mlx5e_change_rep_mtu, + .ndo_change_mtu = mlx5e_uplink_rep_change_mtu, }; static void mlx5e_build_rep_params(struct net_device *netdev) { struct mlx5e_priv *priv = netdev_priv(netdev); + struct mlx5e_rep_priv *rpriv = priv->ppriv; + struct mlx5_eswitch_rep *rep = rpriv->rep; struct mlx5_core_dev *mdev = priv->mdev; struct mlx5e_params *params; @@ -1218,7 +1267,12 @@ static void mlx5e_build_rep_params(struct net_device *netdev) params = &priv->channels.params; params->hard_mtu = MLX5E_ETH_HARD_MTU; params->sw_mtu = netdev->mtu; - params->log_sq_size = MLX5E_REP_PARAMS_LOG_SQ_SIZE; + + /* SQ */ + if (rep->vport == FDB_UPLINK_VPORT) + params->log_sq_size = MLX5E_PARAMS_DEFAULT_LOG_SQ_SIZE; + else + params->log_sq_size = MLX5E_REP_PARAMS_LOG_SQ_SIZE; /* RQ */ mlx5e_build_rq_params(mdev, params); @@ -1238,10 +1292,20 @@ static void mlx5e_build_rep_params(struct net_device *netdev) static void mlx5e_build_rep_netdev(struct net_device *netdev) { struct mlx5e_priv *priv = netdev_priv(netdev); + struct mlx5e_rep_priv *rpriv = priv->ppriv; + struct mlx5_eswitch_rep *rep = rpriv->rep; struct mlx5_core_dev *mdev = priv->mdev; u16 max_mtu; - netdev->netdev_ops = &mlx5e_netdev_ops_rep; + if (rep->vport == FDB_UPLINK_VPORT) { + SET_NETDEV_DEV(netdev, &priv->mdev->pdev->dev); + netdev->netdev_ops = &mlx5e_netdev_ops_uplink_rep; + /* we want a persistent mac for the uplink rep */ + mlx5_query_nic_vport_mac_address(mdev, 0, netdev->dev_addr); + } else { + netdev->netdev_ops = &mlx5e_netdev_ops_vf_rep; + eth_hw_addr_random(netdev); + } netdev->watchdog_timeo = 15 * HZ; @@ -1262,28 +1326,35 @@ static void mlx5e_build_rep_netdev(struct net_device *netdev) netdev->features |= netdev->hw_features; - eth_hw_addr_random(netdev); - netdev->min_mtu = ETH_MIN_MTU; mlx5_query_port_max_mtu(mdev, &max_mtu, 1); netdev->max_mtu = MLX5E_HW2SW_MTU(&priv->channels.params, max_mtu); } +static int mlx5e_rep_get_default_num_channels(struct mlx5_eswitch_rep *rep, + struct net_device *netdev) +{ + if (rep->vport == FDB_UPLINK_VPORT) + return mlx5e_get_netdev_max_channels(netdev); + else + return 1; +} + static int mlx5e_init_rep(struct mlx5_core_dev *mdev, struct net_device *netdev, const struct mlx5e_profile *profile, void *ppriv) { struct mlx5e_priv *priv = netdev_priv(netdev); + struct mlx5e_rep_priv *rpriv = ppriv; int err; err = mlx5e_netdev_init(netdev, priv, mdev, profile, ppriv); if (err) return err; - priv->channels.params.num_channels = - mlx5e_get_netdev_max_channels(netdev); + mlx5e_rep_get_default_num_channels(rpriv->rep, netdev); mlx5e_build_rep_params(netdev); mlx5e_build_rep_netdev(netdev); @@ -1408,14 +1479,60 @@ static void mlx5e_cleanup_rep_rx(struct mlx5e_priv *priv) static int mlx5e_init_rep_tx(struct mlx5e_priv *priv) { - int err; + struct mlx5e_rep_priv *rpriv = priv->ppriv; + struct mlx5_rep_uplink_priv *uplink_priv; + int tc, err; err = mlx5e_create_tises(priv); if (err) { mlx5_core_warn(priv->mdev, "create tises failed, %d\n", err); return err; } + + if (rpriv->rep->vport == FDB_UPLINK_VPORT) { + uplink_priv = &rpriv->uplink_priv; + + /* init shared tc flow table */ + err = mlx5e_tc_esw_init(&uplink_priv->tc_ht); + if (err) + goto destroy_tises; + + /* init indirect block notifications */ + INIT_LIST_HEAD(&uplink_priv->tc_indr_block_priv_list); + uplink_priv->netdevice_nb.notifier_call = mlx5e_nic_rep_netdevice_event; + err = register_netdevice_notifier(&uplink_priv->netdevice_nb); + if (err) { + mlx5_core_err(priv->mdev, "Failed to register netdev notifier\n"); + goto tc_esw_cleanup; + } + } + return 0; + +tc_esw_cleanup: + mlx5e_tc_esw_cleanup(&uplink_priv->tc_ht); +destroy_tises: + for (tc = 0; tc < priv->profile->max_tc; tc++) + mlx5e_destroy_tis(priv->mdev, priv->tisn[tc]); + return err; +} + +static void mlx5e_cleanup_rep_tx(struct mlx5e_priv *priv) +{ + struct mlx5e_rep_priv *rpriv = priv->ppriv; + int tc; + + for (tc = 0; tc < priv->profile->max_tc; tc++) + mlx5e_destroy_tis(priv->mdev, priv->tisn[tc]); + + if (rpriv->rep->vport == FDB_UPLINK_VPORT) { + /* clean indirect TC block notifications */ + unregister_netdevice_notifier(&rpriv->uplink_priv.netdevice_nb); + mlx5e_rep_indr_clean_block_privs(rpriv); + + /* delete shared tc flow table */ + mlx5e_tc_esw_cleanup(&rpriv->uplink_priv.tc_ht); + } } static const struct mlx5e_profile mlx5e_rep_profile = { @@ -1424,7 +1541,7 @@ static const struct mlx5e_profile mlx5e_rep_profile = { .init_rx = mlx5e_init_rep_rx, .cleanup_rx = mlx5e_cleanup_rep_rx, .init_tx = mlx5e_init_rep_tx, - .cleanup_tx = mlx5e_cleanup_nic_tx, + .cleanup_tx = mlx5e_cleanup_rep_tx, .update_stats = mlx5e_rep_update_hw_counters, .update_carrier = NULL, .rx_handlers.handle_rx_cqe = mlx5e_handle_rx_cqe_rep, @@ -1507,6 +1624,9 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) if (!rpriv) return -ENOMEM; + /* rpriv->rep to be looked up when profile->init() is called */ + rpriv->rep = rep; + nch = mlx5e_get_max_num_channels(dev); netdev = mlx5e_create_netdev(dev, &mlx5e_rep_profile, nch, rpriv); if (!netdev) { @@ -1517,7 +1637,6 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) } rpriv->netdev = netdev; - rpriv->rep = rep; rep->rep_if[REP_ETH].priv = rpriv; INIT_LIST_HEAD(&rpriv->vport_sqs_list); @@ -1580,14 +1699,14 @@ static void *mlx5e_vport_rep_get_proto_dev(struct mlx5_eswitch_rep *rep) return rpriv->netdev; } -static void mlx5e_rep_register_vf_vports(struct mlx5e_priv *priv) +void mlx5e_rep_register_vport_reps(struct mlx5e_priv *priv) { struct mlx5_core_dev *mdev = priv->mdev; struct mlx5_eswitch *esw = mdev->priv.eswitch; int total_vfs = MLX5_TOTAL_VPORTS(mdev); int vport; - for (vport = 1; vport < total_vfs; vport++) { + for (vport = 0; vport < total_vfs; vport++) { struct mlx5_eswitch_rep_if rep_if = {}; rep_if.load = mlx5e_vport_rep_load; @@ -1597,46 +1716,17 @@ static void mlx5e_rep_register_vf_vports(struct mlx5e_priv *priv) } } -static void mlx5e_rep_unregister_vf_vports(struct mlx5e_priv *priv) +void mlx5e_rep_unregister_vport_reps(struct mlx5e_priv *priv) { struct mlx5_core_dev *mdev = priv->mdev; struct mlx5_eswitch *esw = mdev->priv.eswitch; int total_vfs = MLX5_TOTAL_VPORTS(mdev); int vport; - for (vport = 1; vport < total_vfs; vport++) + for (vport = total_vfs - 1; vport >= 0; vport--) mlx5_eswitch_unregister_vport_rep(esw, vport, REP_ETH); } -void mlx5e_register_vport_reps(struct mlx5e_priv *priv) -{ - struct mlx5_core_dev *mdev = priv->mdev; - struct mlx5_eswitch *esw = mdev->priv.eswitch; - struct mlx5_eswitch_rep_if rep_if; - struct mlx5e_rep_priv *rpriv; - - rpriv = priv->ppriv; - rpriv->netdev = priv->netdev; - - rep_if.load = mlx5e_nic_rep_load; - rep_if.unload = mlx5e_nic_rep_unload; - rep_if.get_proto_dev = mlx5e_vport_rep_get_proto_dev; - rep_if.priv = rpriv; - INIT_LIST_HEAD(&rpriv->vport_sqs_list); - mlx5_eswitch_register_vport_rep(esw, 0, &rep_if, REP_ETH); /* UPLINK PF vport*/ - - mlx5e_rep_register_vf_vports(priv); /* VFs vports */ -} - -void mlx5e_unregister_vport_reps(struct mlx5e_priv *priv) -{ - struct mlx5_core_dev *mdev = priv->mdev; - struct mlx5_eswitch *esw = mdev->priv.eswitch; - - mlx5e_rep_unregister_vf_vports(priv); /* VFs vports */ - mlx5_eswitch_unregister_vport_rep(esw, 0, REP_ETH); /* UPLINK PF*/ -} - void *mlx5e_alloc_nic_rep_priv(struct mlx5_core_dev *mdev) { struct mlx5_eswitch *esw = mdev->priv.eswitch; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h index 2c8798332c32..07d5964356f1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h @@ -162,8 +162,8 @@ struct mlx5e_rep_sq { }; void *mlx5e_alloc_nic_rep_priv(struct mlx5_core_dev *mdev); -void mlx5e_register_vport_reps(struct mlx5e_priv *priv); -void mlx5e_unregister_vport_reps(struct mlx5e_priv *priv); +void mlx5e_rep_register_vport_reps(struct mlx5e_priv *priv); +void mlx5e_rep_unregister_vport_reps(struct mlx5e_priv *priv); bool mlx5e_is_uplink_rep(struct mlx5e_priv *priv); int mlx5e_add_sqs_fwd_rules(struct mlx5e_priv *priv); void mlx5e_remove_sqs_fwd_rules(struct mlx5e_priv *priv); @@ -171,7 +171,6 @@ void mlx5e_remove_sqs_fwd_rules(struct mlx5e_priv *priv); int mlx5e_get_offload_stats(int attr_id, const struct net_device *dev, void *sp); bool mlx5e_has_offload_stats(const struct net_device *dev, int attr_id); -int mlx5e_attr_get(struct net_device *dev, struct switchdev_attr *attr); void mlx5e_handle_rx_cqe_rep(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe); int mlx5e_rep_encap_entry_attach(struct mlx5e_priv *priv, @@ -181,11 +180,15 @@ void mlx5e_rep_encap_entry_detach(struct mlx5e_priv *priv, void mlx5e_rep_queue_neigh_stats_work(struct mlx5e_priv *priv); #else /* CONFIG_MLX5_ESWITCH */ -static inline void mlx5e_register_vport_reps(struct mlx5e_priv *priv) {} -static inline void mlx5e_unregister_vport_reps(struct mlx5e_priv *priv) {} +static inline void mlx5e_rep_register_vport_reps(struct mlx5e_priv *priv) {} +static inline void mlx5e_rep_unregister_vport_reps(struct mlx5e_priv *priv) {} static inline bool mlx5e_is_uplink_rep(struct mlx5e_priv *priv) { return false; } static inline int mlx5e_add_sqs_fwd_rules(struct mlx5e_priv *priv) { return 0; } static inline void mlx5e_remove_sqs_fwd_rules(struct mlx5e_priv *priv) {} #endif +static inline bool mlx5e_is_vport_rep(struct mlx5e_priv *priv) +{ + return (MLX5_ESWITCH_MANAGER(priv->mdev) && priv->ppriv); +} #endif /* __MLX5E_REP_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c index 8224f1e062a8..3071a44e2f30 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c @@ -486,7 +486,7 @@ static int mlx5e_grp_802_3_fill_stats(struct mlx5e_priv *priv, u64 *data, #define MLX5_BASIC_PPCNT_SUPPORTED(mdev) \ (MLX5_CAP_GEN(mdev, pcam_reg) ? MLX5_CAP_PCAM_REG(mdev, ppcnt) : 1) -static void mlx5e_grp_802_3_update_stats(struct mlx5e_priv *priv) +void mlx5e_grp_802_3_update_stats(struct mlx5e_priv *priv) { struct mlx5e_pport_stats *pstats = &priv->stats.pport; struct mlx5_core_dev *mdev = priv->mdev; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h index 3f8e870ef4c9..807e60582a6b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h @@ -280,5 +280,6 @@ extern const struct mlx5e_stats_grp mlx5e_stats_grps[]; extern const int mlx5e_num_stats_grps; void mlx5e_grp_sw_update_stats(struct mlx5e_priv *priv); +void mlx5e_grp_802_3_update_stats(struct mlx5e_priv *priv); #endif /* __MLX5_EN_STATS_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 53ebb5a48018..c1a9120412b8 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -69,13 +69,13 @@ struct mlx5_nic_flow_attr { enum { MLX5E_TC_FLOW_INGRESS = MLX5E_TC_INGRESS, MLX5E_TC_FLOW_EGRESS = MLX5E_TC_EGRESS, - MLX5E_TC_FLOW_ESWITCH = BIT(MLX5E_TC_FLOW_BASE), - MLX5E_TC_FLOW_NIC = BIT(MLX5E_TC_FLOW_BASE + 1), - MLX5E_TC_FLOW_OFFLOADED = BIT(MLX5E_TC_FLOW_BASE + 2), - MLX5E_TC_FLOW_HAIRPIN = BIT(MLX5E_TC_FLOW_BASE + 3), - MLX5E_TC_FLOW_HAIRPIN_RSS = BIT(MLX5E_TC_FLOW_BASE + 4), - MLX5E_TC_FLOW_SLOW = BIT(MLX5E_TC_FLOW_BASE + 5), - MLX5E_TC_FLOW_DUP = BIT(MLX5E_TC_FLOW_BASE + 6), + MLX5E_TC_FLOW_ESWITCH = MLX5E_TC_ESW_OFFLOAD, + MLX5E_TC_FLOW_NIC = MLX5E_TC_NIC_OFFLOAD, + MLX5E_TC_FLOW_OFFLOADED = BIT(MLX5E_TC_FLOW_BASE), + MLX5E_TC_FLOW_HAIRPIN = BIT(MLX5E_TC_FLOW_BASE + 1), + MLX5E_TC_FLOW_HAIRPIN_RSS = BIT(MLX5E_TC_FLOW_BASE + 2), + MLX5E_TC_FLOW_SLOW = BIT(MLX5E_TC_FLOW_BASE + 3), + MLX5E_TC_FLOW_DUP = BIT(MLX5E_TC_FLOW_BASE + 4), }; #define MLX5E_TC_MAX_SPLITS 1 @@ -834,7 +834,7 @@ static void mlx5e_tc_del_nic_flow(struct mlx5e_priv *priv, mlx5_del_flow_rules(flow->rule[0]); mlx5_fc_destroy(priv->mdev, counter); - if (!mlx5e_tc_num_filters(priv) && priv->fs.tc.t) { + if (!mlx5e_tc_num_filters(priv, MLX5E_TC_NIC_OFFLOAD) && priv->fs.tc.t) { mlx5_destroy_flow_table(priv->fs.tc.t); priv->fs.tc.t = NULL; } @@ -2686,6 +2686,11 @@ static void get_flags(int flags, u16 *flow_flags) if (flags & MLX5E_TC_EGRESS) __flow_flags |= MLX5E_TC_FLOW_EGRESS; + if (flags & MLX5E_TC_ESW_OFFLOAD) + __flow_flags |= MLX5E_TC_FLOW_ESWITCH; + if (flags & MLX5E_TC_NIC_OFFLOAD) + __flow_flags |= MLX5E_TC_FLOW_NIC; + *flow_flags = __flow_flags; } @@ -2696,15 +2701,15 @@ static const struct rhashtable_params tc_ht_params = { .automatic_shrinking = true, }; -static struct rhashtable *get_tc_ht(struct mlx5e_priv *priv) +static struct rhashtable *get_tc_ht(struct mlx5e_priv *priv, int flags) { struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; struct mlx5e_rep_priv *uplink_rpriv; - if (MLX5_VPORT_MANAGER(priv->mdev) && esw->mode == SRIOV_OFFLOADS) { + if (flags & MLX5E_TC_ESW_OFFLOAD) { uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH); return &uplink_rpriv->uplink_priv.tc_ht; - } else + } else /* NIC offload */ return &priv->fs.tc.ht; } @@ -2973,7 +2978,7 @@ int mlx5e_configure_flower(struct net_device *dev, struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, int flags) { struct netlink_ext_ack *extack = f->common.extack; - struct rhashtable *tc_ht = get_tc_ht(priv); + struct rhashtable *tc_ht = get_tc_ht(priv, flags); struct mlx5e_tc_flow *flow; int err = 0; @@ -3018,7 +3023,7 @@ static bool same_flow_direction(struct mlx5e_tc_flow *flow, int flags) int mlx5e_delete_flower(struct net_device *dev, struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, int flags) { - struct rhashtable *tc_ht = get_tc_ht(priv); + struct rhashtable *tc_ht = get_tc_ht(priv, flags); struct mlx5e_tc_flow *flow; flow = rhashtable_lookup_fast(tc_ht, &f->cookie, tc_ht_params); @@ -3038,7 +3043,7 @@ int mlx5e_stats_flower(struct net_device *dev, struct mlx5e_priv *priv, struct tc_cls_flower_offload *f, int flags) { struct mlx5_devcom *devcom = priv->mdev->priv.devcom; - struct rhashtable *tc_ht = get_tc_ht(priv); + struct rhashtable *tc_ht = get_tc_ht(priv, flags); struct mlx5_eswitch *peer_esw; struct mlx5e_tc_flow *flow; struct mlx5_fc *counter; @@ -3168,7 +3173,7 @@ void mlx5e_tc_nic_cleanup(struct mlx5e_priv *priv) if (tc->netdevice_nb.notifier_call) unregister_netdevice_notifier(&tc->netdevice_nb); - rhashtable_free_and_destroy(&tc->ht, _mlx5e_tc_del_flow, NULL); + rhashtable_destroy(&tc->ht); if (!IS_ERR_OR_NULL(tc->t)) { mlx5_destroy_flow_table(tc->t); @@ -3186,9 +3191,9 @@ void mlx5e_tc_esw_cleanup(struct rhashtable *tc_ht) rhashtable_free_and_destroy(tc_ht, _mlx5e_tc_del_flow, NULL); } -int mlx5e_tc_num_filters(struct mlx5e_priv *priv) +int mlx5e_tc_num_filters(struct mlx5e_priv *priv, int flags) { - struct rhashtable *tc_ht = get_tc_ht(priv); + struct rhashtable *tc_ht = get_tc_ht(priv, flags); return atomic_read(&tc_ht->nelems); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h index a15c08a35054..d2d87f978c06 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h @@ -42,7 +42,9 @@ enum { MLX5E_TC_INGRESS = BIT(0), MLX5E_TC_EGRESS = BIT(1), - MLX5E_TC_LAST_EXPORTED_BIT = 1, + MLX5E_TC_NIC_OFFLOAD = BIT(2), + MLX5E_TC_ESW_OFFLOAD = BIT(3), + MLX5E_TC_LAST_EXPORTED_BIT = 3, }; int mlx5e_tc_nic_init(struct mlx5e_priv *priv); @@ -68,13 +70,13 @@ void mlx5e_tc_encap_flows_del(struct mlx5e_priv *priv, struct mlx5e_neigh_hash_entry; void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe); -int mlx5e_tc_num_filters(struct mlx5e_priv *priv); +int mlx5e_tc_num_filters(struct mlx5e_priv *priv, int flags); #else /* CONFIG_MLX5_ESWITCH */ static inline int mlx5e_tc_nic_init(struct mlx5e_priv *priv) { return 0; } static inline void mlx5e_tc_nic_cleanup(struct mlx5e_priv *priv) {} -static inline int mlx5e_tc_num_filters(struct mlx5e_priv *priv) { return 0; } +static inline int mlx5e_tc_num_filters(struct mlx5e_priv *priv, int flags) { return 0; } #endif #endif /* __MLX5_EN_TC_H__ */ -- cgit v1.2.3-55-g7522 From 13e509a4c1949855ddd5ef2e23bafdfb27b8337a Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Tue, 13 Feb 2018 18:14:42 +0200 Subject: net/mlx5e: Remove leftover code from the PF netdev being uplink rep Remove some last leftovers from using the PF netdev as the e-switch uplink representor. Signed-off-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 20 +---- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 105 ++-------------------- drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | 3 - 3 files changed, 9 insertions(+), 119 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index cdd8b5efd96a..9ee50de4de33 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -4381,8 +4381,6 @@ const struct net_device_ops mlx5e_netdev_ops = { .ndo_get_vf_config = mlx5e_get_vf_config, .ndo_set_vf_link_state = mlx5e_set_vf_link_state, .ndo_get_vf_stats = mlx5e_get_vf_stats, - .ndo_has_offload_stats = mlx5e_has_offload_stats, - .ndo_get_offload_stats = mlx5e_get_offload_stats, #endif }; @@ -5132,7 +5130,6 @@ static void mlx5e_detach(struct mlx5_core_dev *mdev, void *vpriv) static void *mlx5e_add(struct mlx5_core_dev *mdev) { struct net_device *netdev; - void *rpriv = NULL; void *priv; int err; int nch; @@ -5141,21 +5138,11 @@ static void *mlx5e_add(struct mlx5_core_dev *mdev) if (err) return NULL; -#ifdef CONFIG_MLX5_ESWITCH - if (MLX5_ESWITCH_MANAGER(mdev)) { - rpriv = mlx5e_alloc_nic_rep_priv(mdev); - if (!rpriv) { - mlx5_core_warn(mdev, "Failed to alloc NIC rep priv data\n"); - return NULL; - } - } -#endif - nch = mlx5e_get_max_num_channels(mdev); - netdev = mlx5e_create_netdev(mdev, &mlx5e_nic_profile, nch, rpriv); + netdev = mlx5e_create_netdev(mdev, &mlx5e_nic_profile, nch, NULL); if (!netdev) { mlx5_core_err(mdev, "mlx5e_create_netdev failed\n"); - goto err_free_rpriv; + return NULL; } priv = netdev_priv(netdev); @@ -5181,8 +5168,6 @@ err_detach: mlx5e_detach(mdev, priv); err_destroy_netdev: mlx5e_destroy_netdev(priv); -err_free_rpriv: - kfree(rpriv); return NULL; } @@ -5197,7 +5182,6 @@ static void mlx5e_remove(struct mlx5_core_dev *mdev, void *vpriv) unregister_netdev(priv->netdev); mlx5e_detach(mdev, vpriv); mlx5e_destroy_netdev(priv); - kfree(ppriv); } static struct mlx5_interface mlx5e_interface = { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index b48ba3234ac5..73c6f24771bd 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1139,22 +1139,7 @@ bool mlx5e_is_uplink_rep(struct mlx5e_priv *priv) return (rep->vport == FDB_UPLINK_VPORT); } -static bool mlx5e_is_vf_vport_rep(struct mlx5e_priv *priv) -{ - struct mlx5e_rep_priv *rpriv = priv->ppriv; - struct mlx5_eswitch_rep *rep; - - if (!MLX5_ESWITCH_MANAGER(priv->mdev)) - return false; - - rep = rpriv->rep; - if (rep && rep->vport != FDB_UPLINK_VPORT) - return true; - - return false; -} - -bool mlx5e_has_offload_stats(const struct net_device *dev, int attr_id) +static bool mlx5e_rep_has_offload_stats(const struct net_device *dev, int attr_id) { switch (attr_id) { case IFLA_OFFLOAD_XSTATS_CPU_HIT: @@ -1183,8 +1168,8 @@ mlx5e_get_sw_stats64(const struct net_device *dev, return 0; } -int mlx5e_get_offload_stats(int attr_id, const struct net_device *dev, - void *sp) +static int mlx5e_rep_get_offload_stats(int attr_id, const struct net_device *dev, + void *sp) { switch (attr_id) { case IFLA_OFFLOAD_XSTATS_CPU_HIT: @@ -1235,8 +1220,8 @@ static const struct net_device_ops mlx5e_netdev_ops_vf_rep = { .ndo_get_phys_port_name = mlx5e_rep_get_phys_port_name, .ndo_setup_tc = mlx5e_rep_setup_tc, .ndo_get_stats64 = mlx5e_vf_rep_get_stats, - .ndo_has_offload_stats = mlx5e_has_offload_stats, - .ndo_get_offload_stats = mlx5e_get_offload_stats, + .ndo_has_offload_stats = mlx5e_rep_has_offload_stats, + .ndo_get_offload_stats = mlx5e_rep_get_offload_stats, .ndo_change_mtu = mlx5e_vf_rep_change_mtu, }; @@ -1247,8 +1232,8 @@ static const struct net_device_ops mlx5e_netdev_ops_uplink_rep = { .ndo_get_phys_port_name = mlx5e_rep_get_phys_port_name, .ndo_setup_tc = mlx5e_rep_setup_tc, .ndo_get_stats64 = mlx5e_get_stats, - .ndo_has_offload_stats = mlx5e_has_offload_stats, - .ndo_get_offload_stats = mlx5e_get_offload_stats, + .ndo_has_offload_stats = mlx5e_rep_has_offload_stats, + .ndo_get_offload_stats = mlx5e_rep_get_offload_stats, .ndo_change_mtu = mlx5e_uplink_rep_change_mtu, }; @@ -1550,69 +1535,6 @@ static const struct mlx5e_profile mlx5e_rep_profile = { }; /* e-Switch vport representors */ - -static int -mlx5e_nic_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) -{ - struct mlx5e_rep_priv *rpriv = mlx5e_rep_to_rep_priv(rep); - struct mlx5e_priv *priv = netdev_priv(rpriv->netdev); - struct mlx5_rep_uplink_priv *uplink_priv = &rpriv->uplink_priv; - int err; - - if (test_bit(MLX5E_STATE_OPENED, &priv->state)) { - err = mlx5e_add_sqs_fwd_rules(priv); - if (err) - return err; - } - - err = mlx5e_rep_neigh_init(rpriv); - if (err) - goto err_remove_sqs; - - /* init shared tc flow table */ - err = mlx5e_tc_esw_init(&uplink_priv->tc_ht); - if (err) - goto err_neigh_cleanup; - - /* init indirect block notifications */ - INIT_LIST_HEAD(&uplink_priv->tc_indr_block_priv_list); - uplink_priv->netdevice_nb.notifier_call = mlx5e_nic_rep_netdevice_event; - err = register_netdevice_notifier(&uplink_priv->netdevice_nb); - if (err) { - mlx5_core_err(priv->mdev, "Failed to register netdev notifier\n"); - goto err_indirect_block_cleanup; - } - - return 0; - -err_indirect_block_cleanup: - mlx5e_tc_esw_cleanup(&uplink_priv->tc_ht); -err_neigh_cleanup: - mlx5e_rep_neigh_cleanup(rpriv); -err_remove_sqs: - mlx5e_remove_sqs_fwd_rules(priv); - return err; -} - -static void -mlx5e_nic_rep_unload(struct mlx5_eswitch_rep *rep) -{ - struct mlx5e_rep_priv *rpriv = mlx5e_rep_to_rep_priv(rep); - struct mlx5e_priv *priv = netdev_priv(rpriv->netdev); - - if (test_bit(MLX5E_STATE_OPENED, &priv->state)) - mlx5e_remove_sqs_fwd_rules(priv); - - /* clean indirect TC block notifications */ - unregister_netdevice_notifier(&rpriv->uplink_priv.netdevice_nb); - mlx5e_rep_indr_clean_block_privs(rpriv); - - /* clean uplink offloaded TC rules, delete shared tc flow table */ - mlx5e_tc_esw_cleanup(&rpriv->uplink_priv.tc_ht); - - mlx5e_rep_neigh_cleanup(rpriv); -} - static int mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) { @@ -1726,16 +1648,3 @@ void mlx5e_rep_unregister_vport_reps(struct mlx5e_priv *priv) for (vport = total_vfs - 1; vport >= 0; vport--) mlx5_eswitch_unregister_vport_rep(esw, vport, REP_ETH); } - -void *mlx5e_alloc_nic_rep_priv(struct mlx5_core_dev *mdev) -{ - struct mlx5_eswitch *esw = mdev->priv.eswitch; - struct mlx5e_rep_priv *rpriv; - - rpriv = kzalloc(sizeof(*rpriv), GFP_KERNEL); - if (!rpriv) - return NULL; - - rpriv->rep = &esw->offloads.vport_reps[0]; - return rpriv; -} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h index 07d5964356f1..bb2d2484eb64 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h @@ -168,9 +168,6 @@ bool mlx5e_is_uplink_rep(struct mlx5e_priv *priv); int mlx5e_add_sqs_fwd_rules(struct mlx5e_priv *priv); void mlx5e_remove_sqs_fwd_rules(struct mlx5e_priv *priv); -int mlx5e_get_offload_stats(int attr_id, const struct net_device *dev, void *sp); -bool mlx5e_has_offload_stats(const struct net_device *dev, int attr_id); - void mlx5e_handle_rx_cqe_rep(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe); int mlx5e_rep_encap_entry_attach(struct mlx5e_priv *priv, -- cgit v1.2.3-55-g7522 From aec002f6f82cbfb39f92605614a88dd7a200a766 Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Wed, 7 Nov 2018 16:34:52 +0200 Subject: net/mlx5e: Uninstantiate esw manager vport netdev on switchdev mode Now, when we have a dedicated uplink representor, the netdev instance set over the esw manager vport (PF) is of no-use. As such, remove it once we're on switchdev mode and get it back to life when off switchdev. This is done by reloading the Ethernet interface as well (we already do that for the IB interface) from the eswitch code while going in/out of switchdev mode. The Eth add/remove entries are modified to act differently when called in switchdev mode. In this case we only deal with registration of the eth vport representors. The rep netdevices are created from the eswitch call to load the registered eth representors. Signed-off-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 24 +++++++++++++++-------- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 22 +++++++++++++++------ drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | 6 ++---- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 11 ++++++++--- 4 files changed, 42 insertions(+), 21 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 9ee50de4de33..0fd5752fb181 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -4894,9 +4894,6 @@ static void mlx5e_nic_enable(struct mlx5e_priv *priv) if (mlx5e_monitor_counter_supported(priv)) mlx5e_monitor_counter_init(priv); - if (MLX5_ESWITCH_MANAGER(priv->mdev)) - mlx5e_rep_register_vport_reps(priv); - if (netdev->reg_state != NETREG_REGISTERED) return; #ifdef CONFIG_MLX5_CORE_EN_DCB @@ -4929,9 +4926,6 @@ static void mlx5e_nic_disable(struct mlx5e_priv *priv) queue_work(priv->wq, &priv->set_rx_mode_work); - if (MLX5_ESWITCH_MANAGER(priv->mdev)) - mlx5e_rep_unregister_vport_reps(priv); - if (mlx5e_monitor_counter_supported(priv)) mlx5e_monitor_counter_cleanup(priv); @@ -5138,6 +5132,14 @@ static void *mlx5e_add(struct mlx5_core_dev *mdev) if (err) return NULL; +#ifdef CONFIG_MLX5_ESWITCH + if (MLX5_ESWITCH_MANAGER(mdev) && + mlx5_eswitch_mode(mdev->priv.eswitch) == SRIOV_OFFLOADS) { + mlx5e_rep_register_vport_reps(mdev); + return mdev; + } +#endif + nch = mlx5e_get_max_num_channels(mdev); netdev = mlx5e_create_netdev(mdev, &mlx5e_nic_profile, nch, NULL); if (!netdev) { @@ -5173,9 +5175,15 @@ err_destroy_netdev: static void mlx5e_remove(struct mlx5_core_dev *mdev, void *vpriv) { - struct mlx5e_priv *priv = vpriv; - void *ppriv = priv->ppriv; + struct mlx5e_priv *priv; +#ifdef CONFIG_MLX5_ESWITCH + if (MLX5_ESWITCH_MANAGER(mdev) && vpriv == mdev) { + mlx5e_rep_unregister_vport_reps(mdev); + return; + } +#endif + priv = vpriv; #ifdef CONFIG_MLX5_CORE_EN_DCB mlx5e_dcbnl_delete_app(priv); #endif diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index 73c6f24771bd..e4f8f639dea0 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1562,11 +1562,17 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) rep->rep_if[REP_ETH].priv = rpriv; INIT_LIST_HEAD(&rpriv->vport_sqs_list); + if (rep->vport == FDB_UPLINK_VPORT) { + err = mlx5e_create_mdev_resources(dev); + if (err) + goto err_destroy_netdev; + } + err = mlx5e_attach_netdev(netdev_priv(netdev)); if (err) { pr_warn("Failed to attach representor netdev for vport %d\n", rep->vport); - goto err_destroy_netdev; + goto err_destroy_mdev_resources; } err = mlx5e_rep_neigh_init(rpriv); @@ -1591,6 +1597,10 @@ err_neigh_cleanup: err_detach_netdev: mlx5e_detach_netdev(netdev_priv(netdev)); +err_destroy_mdev_resources: + if (rep->vport == FDB_UPLINK_VPORT) + mlx5e_destroy_mdev_resources(dev); + err_destroy_netdev: mlx5e_destroy_netdev(netdev_priv(netdev)); kfree(rpriv); @@ -1608,6 +1618,8 @@ mlx5e_vport_rep_unload(struct mlx5_eswitch_rep *rep) unregister_netdev(netdev); mlx5e_rep_neigh_cleanup(rpriv); mlx5e_detach_netdev(priv); + if (rep->vport == FDB_UPLINK_VPORT) + mlx5e_destroy_mdev_resources(priv->mdev); mlx5e_destroy_netdev(priv); kfree(ppriv); /* mlx5e_rep_priv */ } @@ -1621,10 +1633,9 @@ static void *mlx5e_vport_rep_get_proto_dev(struct mlx5_eswitch_rep *rep) return rpriv->netdev; } -void mlx5e_rep_register_vport_reps(struct mlx5e_priv *priv) +void mlx5e_rep_register_vport_reps(struct mlx5_core_dev *mdev) { - struct mlx5_core_dev *mdev = priv->mdev; - struct mlx5_eswitch *esw = mdev->priv.eswitch; + struct mlx5_eswitch *esw = mdev->priv.eswitch; int total_vfs = MLX5_TOTAL_VPORTS(mdev); int vport; @@ -1638,9 +1649,8 @@ void mlx5e_rep_register_vport_reps(struct mlx5e_priv *priv) } } -void mlx5e_rep_unregister_vport_reps(struct mlx5e_priv *priv) +void mlx5e_rep_unregister_vport_reps(struct mlx5_core_dev *mdev) { - struct mlx5_core_dev *mdev = priv->mdev; struct mlx5_eswitch *esw = mdev->priv.eswitch; int total_vfs = MLX5_TOTAL_VPORTS(mdev); int vport; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h index bb2d2484eb64..5645d3cef1bb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h @@ -162,8 +162,8 @@ struct mlx5e_rep_sq { }; void *mlx5e_alloc_nic_rep_priv(struct mlx5_core_dev *mdev); -void mlx5e_rep_register_vport_reps(struct mlx5e_priv *priv); -void mlx5e_rep_unregister_vport_reps(struct mlx5e_priv *priv); +void mlx5e_rep_register_vport_reps(struct mlx5_core_dev *mdev); +void mlx5e_rep_unregister_vport_reps(struct mlx5_core_dev *mdev); bool mlx5e_is_uplink_rep(struct mlx5e_priv *priv); int mlx5e_add_sqs_fwd_rules(struct mlx5e_priv *priv); void mlx5e_remove_sqs_fwd_rules(struct mlx5e_priv *priv); @@ -177,8 +177,6 @@ void mlx5e_rep_encap_entry_detach(struct mlx5e_priv *priv, void mlx5e_rep_queue_neigh_stats_work(struct mlx5e_priv *priv); #else /* CONFIG_MLX5_ESWITCH */ -static inline void mlx5e_rep_register_vport_reps(struct mlx5e_priv *priv) {} -static inline void mlx5e_rep_unregister_vport_reps(struct mlx5e_priv *priv) {} static inline bool mlx5e_is_uplink_rep(struct mlx5e_priv *priv) { return false; } static inline int mlx5e_add_sqs_fwd_rules(struct mlx5e_priv *priv) { return 0; } static inline void mlx5e_remove_sqs_fwd_rules(struct mlx5e_priv *priv) {} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index ef2bec50423d..a44ea7b85614 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -1630,6 +1630,7 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode) esw_warn(esw->dev, "E-Switch engress ACL is not supported by FW\n"); esw_info(esw->dev, "E-Switch enable SRIOV: nvfs(%d) mode (%d)\n", nvfs, mode); + esw->mode = mode; mlx5_lag_update(esw->dev); @@ -1637,8 +1638,8 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode) if (mode == SRIOV_LEGACY) { err = esw_create_legacy_fdb_table(esw); } else { + mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_ETH); mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB); - err = esw_offloads_init(esw, nvfs + 1); } @@ -1669,8 +1670,10 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode) abort: esw->mode = SRIOV_NONE; - if (mode == SRIOV_OFFLOADS) + if (mode == SRIOV_OFFLOADS) { mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB); + mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_ETH); + } return err; } @@ -1712,8 +1715,10 @@ void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw) mlx5_lag_update(esw->dev); - if (old_mode == SRIOV_OFFLOADS) + if (old_mode == SRIOV_OFFLOADS) { mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB); + mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_ETH); + } } int mlx5_eswitch_init(struct mlx5_core_dev *dev) -- cgit v1.2.3-55-g7522 From b36cdb42ad4b114d4b9bbba8770d7a4b8e1a62f6 Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Thu, 8 Nov 2018 20:42:55 +0200 Subject: net/mlx5e: Handle port mtu/link, dcb and lag for uplink reps Take care of setup/teardown for the port link, dcb, lag as well as dealing with port mtu and carrier for e-switch uplink representors. This is achieved by adding a dedicated profile instance for uplink representors which includes the enable/disable and more profile routines which are invoked by the general mlx5e code for netdev attach/detach. Signed-off-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 1 + drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 119 ++++++++++++++++++---- 3 files changed, 102 insertions(+), 20 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 31deafecf679..71c0646c8108 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -940,6 +940,7 @@ int mlx5e_create_tis(struct mlx5_core_dev *mdev, int tc, void mlx5e_destroy_tis(struct mlx5_core_dev *mdev, u32 tisn); int mlx5e_create_tises(struct mlx5e_priv *priv); +void mlx5e_update_carrier(struct mlx5e_priv *priv); int mlx5e_close(struct net_device *netdev); int mlx5e_open(struct net_device *netdev); void mlx5e_update_ndo_stats(struct mlx5e_priv *priv); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 0fd5752fb181..0af61ef3cebb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -225,7 +225,7 @@ void mlx5e_set_rq_type(struct mlx5_core_dev *mdev, struct mlx5e_params *params) MLX5_WQ_TYPE_CYCLIC; } -static void mlx5e_update_carrier(struct mlx5e_priv *priv) +void mlx5e_update_carrier(struct mlx5e_priv *priv) { struct mlx5_core_dev *mdev = priv->mdev; u8 port_state; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index e4f8f639dea0..e676660826fc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1194,19 +1194,20 @@ static int mlx5e_vf_rep_change_mtu(struct net_device *netdev, int new_mtu) return mlx5e_change_mtu(netdev, new_mtu, NULL); } -static int mlx5e_uplink_rep_open(struct net_device *dev) +static int mlx5e_uplink_rep_change_mtu(struct net_device *netdev, int new_mtu) { - int err; - - err = mlx5e_open(dev); - if (!err) - netif_carrier_on(dev); - return err; + return mlx5e_change_mtu(netdev, new_mtu, mlx5e_set_dev_port_mtu); } -static int mlx5e_uplink_rep_change_mtu(struct net_device *netdev, int new_mtu) +static int mlx5e_uplink_rep_set_mac(struct net_device *netdev, void *addr) { - return mlx5e_change_mtu(netdev, new_mtu, mlx5e_set_dev_port_mtu); + struct sockaddr *saddr = addr; + + if (!is_valid_ether_addr(saddr->sa_data)) + return -EADDRNOTAVAIL; + + ether_addr_copy(netdev->dev_addr, saddr->sa_data); + return 0; } static const struct switchdev_ops mlx5e_rep_switchdev_ops = { @@ -1226,9 +1227,10 @@ static const struct net_device_ops mlx5e_netdev_ops_vf_rep = { }; static const struct net_device_ops mlx5e_netdev_ops_uplink_rep = { - .ndo_open = mlx5e_uplink_rep_open, + .ndo_open = mlx5e_open, .ndo_stop = mlx5e_close, .ndo_start_xmit = mlx5e_xmit, + .ndo_set_mac_address = mlx5e_uplink_rep_set_mac, .ndo_get_phys_port_name = mlx5e_rep_get_phys_port_name, .ndo_setup_tc = mlx5e_rep_setup_tc, .ndo_get_stats64 = mlx5e_get_stats, @@ -1280,13 +1282,16 @@ static void mlx5e_build_rep_netdev(struct net_device *netdev) struct mlx5e_rep_priv *rpriv = priv->ppriv; struct mlx5_eswitch_rep *rep = rpriv->rep; struct mlx5_core_dev *mdev = priv->mdev; - u16 max_mtu; if (rep->vport == FDB_UPLINK_VPORT) { SET_NETDEV_DEV(netdev, &priv->mdev->pdev->dev); netdev->netdev_ops = &mlx5e_netdev_ops_uplink_rep; /* we want a persistent mac for the uplink rep */ mlx5_query_nic_vport_mac_address(mdev, 0, netdev->dev_addr); +#ifdef CONFIG_MLX5_CORE_EN_DCB + if (MLX5_CAP_GEN(mdev, qos)) + netdev->dcbnl_ops = &mlx5e_dcbnl_ops; +#endif } else { netdev->netdev_ops = &mlx5e_netdev_ops_vf_rep; eth_hw_addr_random(netdev); @@ -1310,10 +1315,6 @@ static void mlx5e_build_rep_netdev(struct net_device *netdev) netdev->hw_features |= NETIF_F_RXCSUM; netdev->features |= netdev->hw_features; - - netdev->min_mtu = ETH_MIN_MTU; - mlx5_query_port_max_mtu(mdev, &max_mtu, 1); - netdev->max_mtu = MLX5E_HW2SW_MTU(&priv->channels.params, max_mtu); } static int mlx5e_rep_get_default_num_channels(struct mlx5_eswitch_rep *rep, @@ -1520,24 +1521,103 @@ static void mlx5e_cleanup_rep_tx(struct mlx5e_priv *priv) } } -static const struct mlx5e_profile mlx5e_rep_profile = { +static void mlx5e_vf_rep_enable(struct mlx5e_priv *priv) +{ + struct net_device *netdev = priv->netdev; + struct mlx5_core_dev *mdev = priv->mdev; + u16 max_mtu; + + netdev->min_mtu = ETH_MIN_MTU; + mlx5_query_port_max_mtu(mdev, &max_mtu, 1); + netdev->max_mtu = MLX5E_HW2SW_MTU(&priv->channels.params, max_mtu); +} + +static int uplink_rep_async_event(struct notifier_block *nb, unsigned long event, void *data) +{ + struct mlx5e_priv *priv = container_of(nb, struct mlx5e_priv, events_nb); + struct mlx5_eqe *eqe = data; + + if (event != MLX5_EVENT_TYPE_PORT_CHANGE) + return NOTIFY_DONE; + + switch (eqe->sub_type) { + case MLX5_PORT_CHANGE_SUBTYPE_DOWN: + case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE: + queue_work(priv->wq, &priv->update_carrier_work); + break; + default: + return NOTIFY_DONE; + } + + return NOTIFY_OK; +} + +static void mlx5e_uplink_rep_enable(struct mlx5e_priv *priv) +{ + struct net_device *netdev = priv->netdev; + struct mlx5_core_dev *mdev = priv->mdev; + u16 max_mtu; + + netdev->min_mtu = ETH_MIN_MTU; + mlx5_query_port_max_mtu(priv->mdev, &max_mtu, 1); + netdev->max_mtu = MLX5E_HW2SW_MTU(&priv->channels.params, max_mtu); + mlx5e_set_dev_port_mtu(priv); + + mlx5_lag_add(mdev, netdev); + priv->events_nb.notifier_call = uplink_rep_async_event; + mlx5_notifier_register(mdev, &priv->events_nb); +#ifdef CONFIG_MLX5_CORE_EN_DCB + mlx5e_dcbnl_initialize(priv); + mlx5e_dcbnl_init_app(priv); +#endif +} + +static void mlx5e_uplink_rep_disable(struct mlx5e_priv *priv) +{ + struct mlx5_core_dev *mdev = priv->mdev; + +#ifdef CONFIG_MLX5_CORE_EN_DCB + mlx5e_dcbnl_delete_app(priv); +#endif + mlx5_notifier_unregister(mdev, &priv->events_nb); + mlx5_lag_remove(mdev); +} + +static const struct mlx5e_profile mlx5e_vf_rep_profile = { .init = mlx5e_init_rep, .cleanup = mlx5e_cleanup_rep, .init_rx = mlx5e_init_rep_rx, .cleanup_rx = mlx5e_cleanup_rep_rx, .init_tx = mlx5e_init_rep_tx, .cleanup_tx = mlx5e_cleanup_rep_tx, - .update_stats = mlx5e_rep_update_hw_counters, - .update_carrier = NULL, + .enable = mlx5e_vf_rep_enable, + .update_stats = mlx5e_vf_rep_update_hw_counters, .rx_handlers.handle_rx_cqe = mlx5e_handle_rx_cqe_rep, .rx_handlers.handle_rx_cqe_mpwqe = mlx5e_handle_rx_cqe_mpwrq, .max_tc = 1, }; +static const struct mlx5e_profile mlx5e_uplink_rep_profile = { + .init = mlx5e_init_rep, + .cleanup = mlx5e_cleanup_rep, + .init_rx = mlx5e_init_rep_rx, + .cleanup_rx = mlx5e_cleanup_rep_rx, + .init_tx = mlx5e_init_rep_tx, + .cleanup_tx = mlx5e_cleanup_rep_tx, + .enable = mlx5e_uplink_rep_enable, + .disable = mlx5e_uplink_rep_disable, + .update_stats = mlx5e_uplink_rep_update_hw_counters, + .update_carrier = mlx5e_update_carrier, + .rx_handlers.handle_rx_cqe = mlx5e_handle_rx_cqe_rep, + .rx_handlers.handle_rx_cqe_mpwqe = mlx5e_handle_rx_cqe_mpwrq, + .max_tc = MLX5E_MAX_NUM_TC, +}; + /* e-Switch vport representors */ static int mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) { + const struct mlx5e_profile *profile; struct mlx5e_rep_priv *rpriv; struct net_device *netdev; int nch, err; @@ -1550,7 +1630,8 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep) rpriv->rep = rep; nch = mlx5e_get_max_num_channels(dev); - netdev = mlx5e_create_netdev(dev, &mlx5e_rep_profile, nch, rpriv); + profile = (rep->vport == FDB_UPLINK_VPORT) ? &mlx5e_uplink_rep_profile : &mlx5e_vf_rep_profile; + netdev = mlx5e_create_netdev(dev, profile, nch, rpriv); if (!netdev) { pr_warn("Failed to create representor netdev for vport %d\n", rep->vport); -- cgit v1.2.3-55-g7522 From 073caf508822750cde02086a6ac1b9131e69692c Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Thu, 1 Nov 2018 19:14:21 +0200 Subject: net/mlx5e: Add sriov and udp tunnel ndo support for the uplink rep Some of the sriov ndo calls are needed also on the switchdev mode - e.g setup VF mac and reading vport stats. Add them to the uplink rep netdev ops. Same for the UDP tunnel ones, need them there to identify offloaded udp tunnel ports. Signed-off-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 12 +++++++++++ drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 26 +++++++++++------------ drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 7 ++++++ 3 files changed, 31 insertions(+), 14 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 71c0646c8108..16fe9e9a14c6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -1004,4 +1004,16 @@ void mlx5e_build_rss_params(struct mlx5e_rss_params *rss_params, u8 mlx5e_params_calculate_tx_min_inline(struct mlx5_core_dev *mdev); void mlx5e_rx_dim_work(struct work_struct *work); void mlx5e_tx_dim_work(struct work_struct *work); + +void mlx5e_add_vxlan_port(struct net_device *netdev, struct udp_tunnel_info *ti); +void mlx5e_del_vxlan_port(struct net_device *netdev, struct udp_tunnel_info *ti); +netdev_features_t mlx5e_features_check(struct sk_buff *skb, + struct net_device *netdev, + netdev_features_t features); +#ifdef CONFIG_MLX5_ESWITCH +int mlx5e_set_vf_mac(struct net_device *dev, int vf, u8 *mac); +int mlx5e_set_vf_rate(struct net_device *dev, int vf, int min_tx_rate, int max_tx_rate); +int mlx5e_get_vf_config(struct net_device *dev, int vf, struct ifla_vf_info *ivi); +int mlx5e_get_vf_stats(struct net_device *dev, int vf, struct ifla_vf_stats *vf_stats); +#endif #endif /* __MLX5_EN_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 0af61ef3cebb..cd6872a9e4dc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -3899,7 +3899,7 @@ static int mlx5e_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) } #ifdef CONFIG_MLX5_ESWITCH -static int mlx5e_set_vf_mac(struct net_device *dev, int vf, u8 *mac) +int mlx5e_set_vf_mac(struct net_device *dev, int vf, u8 *mac) { struct mlx5e_priv *priv = netdev_priv(dev); struct mlx5_core_dev *mdev = priv->mdev; @@ -3936,8 +3936,8 @@ static int mlx5e_set_vf_trust(struct net_device *dev, int vf, bool setting) return mlx5_eswitch_set_vport_trust(mdev->priv.eswitch, vf + 1, setting); } -static int mlx5e_set_vf_rate(struct net_device *dev, int vf, int min_tx_rate, - int max_tx_rate) +int mlx5e_set_vf_rate(struct net_device *dev, int vf, int min_tx_rate, + int max_tx_rate) { struct mlx5e_priv *priv = netdev_priv(dev); struct mlx5_core_dev *mdev = priv->mdev; @@ -3978,8 +3978,8 @@ static int mlx5e_set_vf_link_state(struct net_device *dev, int vf, mlx5_ifla_link2vport(link_state)); } -static int mlx5e_get_vf_config(struct net_device *dev, - int vf, struct ifla_vf_info *ivi) +int mlx5e_get_vf_config(struct net_device *dev, + int vf, struct ifla_vf_info *ivi) { struct mlx5e_priv *priv = netdev_priv(dev); struct mlx5_core_dev *mdev = priv->mdev; @@ -3992,8 +3992,8 @@ static int mlx5e_get_vf_config(struct net_device *dev, return 0; } -static int mlx5e_get_vf_stats(struct net_device *dev, - int vf, struct ifla_vf_stats *vf_stats) +int mlx5e_get_vf_stats(struct net_device *dev, + int vf, struct ifla_vf_stats *vf_stats) { struct mlx5e_priv *priv = netdev_priv(dev); struct mlx5_core_dev *mdev = priv->mdev; @@ -4054,8 +4054,7 @@ static void mlx5e_vxlan_queue_work(struct mlx5e_priv *priv, u16 port, int add) queue_work(priv->wq, &vxlan_work->work); } -static void mlx5e_add_vxlan_port(struct net_device *netdev, - struct udp_tunnel_info *ti) +void mlx5e_add_vxlan_port(struct net_device *netdev, struct udp_tunnel_info *ti) { struct mlx5e_priv *priv = netdev_priv(netdev); @@ -4068,8 +4067,7 @@ static void mlx5e_add_vxlan_port(struct net_device *netdev, mlx5e_vxlan_queue_work(priv, be16_to_cpu(ti->port), 1); } -static void mlx5e_del_vxlan_port(struct net_device *netdev, - struct udp_tunnel_info *ti) +void mlx5e_del_vxlan_port(struct net_device *netdev, struct udp_tunnel_info *ti) { struct mlx5e_priv *priv = netdev_priv(netdev); @@ -4119,9 +4117,9 @@ out: return features & ~(NETIF_F_CSUM_MASK | NETIF_F_GSO_MASK); } -static netdev_features_t mlx5e_features_check(struct sk_buff *skb, - struct net_device *netdev, - netdev_features_t features) +netdev_features_t mlx5e_features_check(struct sk_buff *skb, + struct net_device *netdev, + netdev_features_t features) { struct mlx5e_priv *priv = netdev_priv(netdev); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index e676660826fc..f557aed54564 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1237,6 +1237,13 @@ static const struct net_device_ops mlx5e_netdev_ops_uplink_rep = { .ndo_has_offload_stats = mlx5e_rep_has_offload_stats, .ndo_get_offload_stats = mlx5e_rep_get_offload_stats, .ndo_change_mtu = mlx5e_uplink_rep_change_mtu, + .ndo_udp_tunnel_add = mlx5e_add_vxlan_port, + .ndo_udp_tunnel_del = mlx5e_del_vxlan_port, + .ndo_features_check = mlx5e_features_check, + .ndo_set_vf_mac = mlx5e_set_vf_mac, + .ndo_set_vf_rate = mlx5e_set_vf_rate, + .ndo_get_vf_config = mlx5e_get_vf_config, + .ndo_get_vf_stats = mlx5e_get_vf_stats, }; static void mlx5e_build_rep_params(struct net_device *netdev) -- cgit v1.2.3-55-g7522 From 371289b61ac1ac890c7e98b1bfd3b8f32e38efae Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Tue, 6 Nov 2018 19:31:10 +0200 Subject: net/mlx5e: Expose ethtool pause and link functions to mlx5e callers Towards supporting set/get of global pause for the port and get of the port link ksetting from the uplink representor, expose the relevant entries to other mlx5 callers. Signed-off-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 8 +++ .../net/ethernet/mellanox/mlx5/core/en_ethtool.c | 68 +++++++++++++++------- 2 files changed, 56 insertions(+), 20 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 16fe9e9a14c6..72cc204f9a34 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -973,12 +973,20 @@ int mlx5e_ethtool_get_coalesce(struct mlx5e_priv *priv, struct ethtool_coalesce *coal); int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv, struct ethtool_coalesce *coal); +int mlx5e_ethtool_get_link_ksettings(struct mlx5e_priv *priv, + struct ethtool_link_ksettings *link_ksettings); +int mlx5e_ethtool_set_link_ksettings(struct mlx5e_priv *priv, + const struct ethtool_link_ksettings *link_ksettings); u32 mlx5e_ethtool_get_rxfh_key_size(struct mlx5e_priv *priv); u32 mlx5e_ethtool_get_rxfh_indir_size(struct mlx5e_priv *priv); int mlx5e_ethtool_get_ts_info(struct mlx5e_priv *priv, struct ethtool_ts_info *info); int mlx5e_ethtool_flash_device(struct mlx5e_priv *priv, struct ethtool_flash *flash); +void mlx5e_ethtool_get_pauseparam(struct mlx5e_priv *priv, + struct ethtool_pauseparam *pauseparam); +int mlx5e_ethtool_set_pauseparam(struct mlx5e_priv *priv, + struct ethtool_pauseparam *pauseparam); /* mlx5e generic netdev management API */ int mlx5e_netdev_init(struct net_device *netdev, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c index da4bb62ac528..6af587cf147f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c @@ -787,10 +787,9 @@ static void get_lp_advertising(u32 eth_proto_lp, ptys2ethtool_adver_link(lp_advertising, eth_proto_lp); } -static int mlx5e_get_link_ksettings(struct net_device *netdev, - struct ethtool_link_ksettings *link_ksettings) +int mlx5e_ethtool_get_link_ksettings(struct mlx5e_priv *priv, + struct ethtool_link_ksettings *link_ksettings) { - struct mlx5e_priv *priv = netdev_priv(netdev); struct mlx5_core_dev *mdev = priv->mdev; u32 out[MLX5_ST_SZ_DW(ptys_reg)] = {0}; u32 rx_pause = 0; @@ -806,7 +805,7 @@ static int mlx5e_get_link_ksettings(struct net_device *netdev, err = mlx5_query_port_ptys(mdev, out, sizeof(out), MLX5_PTYS_EN, 1); if (err) { - netdev_err(netdev, "%s: query port ptys failed: %d\n", + netdev_err(priv->netdev, "%s: query port ptys failed: %d\n", __func__, err); goto err_query_regs; } @@ -826,7 +825,7 @@ static int mlx5e_get_link_ksettings(struct net_device *netdev, get_supported(eth_proto_cap, link_ksettings); get_advertising(eth_proto_admin, tx_pause, rx_pause, link_ksettings); - get_speed_duplex(netdev, eth_proto_oper, link_ksettings); + get_speed_duplex(priv->netdev, eth_proto_oper, link_ksettings); eth_proto_oper = eth_proto_oper ? eth_proto_oper : eth_proto_cap; @@ -846,7 +845,7 @@ static int mlx5e_get_link_ksettings(struct net_device *netdev, Autoneg); if (get_fec_supported_advertised(mdev, link_ksettings)) - netdev_dbg(netdev, "%s: FEC caps query failed: %d\n", + netdev_dbg(priv->netdev, "%s: FEC caps query failed: %d\n", __func__, err); if (!an_disable_admin) @@ -857,6 +856,14 @@ err_query_regs: return err; } +static int mlx5e_get_link_ksettings(struct net_device *netdev, + struct ethtool_link_ksettings *link_ksettings) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + + return mlx5e_ethtool_get_link_ksettings(priv, link_ksettings); +} + static u32 mlx5e_ethtool2ptys_adver_link(const unsigned long *link_modes) { u32 i, ptys_modes = 0; @@ -871,10 +878,9 @@ static u32 mlx5e_ethtool2ptys_adver_link(const unsigned long *link_modes) return ptys_modes; } -static int mlx5e_set_link_ksettings(struct net_device *netdev, - const struct ethtool_link_ksettings *link_ksettings) +int mlx5e_ethtool_set_link_ksettings(struct mlx5e_priv *priv, + const struct ethtool_link_ksettings *link_ksettings) { - struct mlx5e_priv *priv = netdev_priv(netdev); struct mlx5_core_dev *mdev = priv->mdev; u32 eth_proto_cap, eth_proto_admin; bool an_changes = false; @@ -894,14 +900,14 @@ static int mlx5e_set_link_ksettings(struct net_device *netdev, err = mlx5_query_port_proto_cap(mdev, ð_proto_cap, MLX5_PTYS_EN); if (err) { - netdev_err(netdev, "%s: query port eth proto cap failed: %d\n", + netdev_err(priv->netdev, "%s: query port eth proto cap failed: %d\n", __func__, err); goto out; } link_modes = link_modes & eth_proto_cap; if (!link_modes) { - netdev_err(netdev, "%s: Not supported link mode(s) requested", + netdev_err(priv->netdev, "%s: Not supported link mode(s) requested", __func__); err = -EINVAL; goto out; @@ -909,7 +915,7 @@ static int mlx5e_set_link_ksettings(struct net_device *netdev, err = mlx5_query_port_proto_admin(mdev, ð_proto_admin, MLX5_PTYS_EN); if (err) { - netdev_err(netdev, "%s: query port eth proto admin failed: %d\n", + netdev_err(priv->netdev, "%s: query port eth proto admin failed: %d\n", __func__, err); goto out; } @@ -931,6 +937,14 @@ out: return err; } +static int mlx5e_set_link_ksettings(struct net_device *netdev, + const struct ethtool_link_ksettings *link_ksettings) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + + return mlx5e_ethtool_set_link_ksettings(priv, link_ksettings); +} + u32 mlx5e_ethtool_get_rxfh_key_size(struct mlx5e_priv *priv) { return sizeof(priv->rss_params.toeplitz_hash_key); @@ -1127,25 +1141,31 @@ static int mlx5e_set_tunable(struct net_device *dev, return err; } -static void mlx5e_get_pauseparam(struct net_device *netdev, - struct ethtool_pauseparam *pauseparam) +void mlx5e_ethtool_get_pauseparam(struct mlx5e_priv *priv, + struct ethtool_pauseparam *pauseparam) { - struct mlx5e_priv *priv = netdev_priv(netdev); struct mlx5_core_dev *mdev = priv->mdev; int err; err = mlx5_query_port_pause(mdev, &pauseparam->rx_pause, &pauseparam->tx_pause); if (err) { - netdev_err(netdev, "%s: mlx5_query_port_pause failed:0x%x\n", + netdev_err(priv->netdev, "%s: mlx5_query_port_pause failed:0x%x\n", __func__, err); } } -static int mlx5e_set_pauseparam(struct net_device *netdev, - struct ethtool_pauseparam *pauseparam) +static void mlx5e_get_pauseparam(struct net_device *netdev, + struct ethtool_pauseparam *pauseparam) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + + mlx5e_ethtool_get_pauseparam(priv, pauseparam); +} + +int mlx5e_ethtool_set_pauseparam(struct mlx5e_priv *priv, + struct ethtool_pauseparam *pauseparam) { - struct mlx5e_priv *priv = netdev_priv(netdev); struct mlx5_core_dev *mdev = priv->mdev; int err; @@ -1156,13 +1176,21 @@ static int mlx5e_set_pauseparam(struct net_device *netdev, pauseparam->rx_pause ? 1 : 0, pauseparam->tx_pause ? 1 : 0); if (err) { - netdev_err(netdev, "%s: mlx5_set_port_pause failed:0x%x\n", + netdev_err(priv->netdev, "%s: mlx5_set_port_pause failed:0x%x\n", __func__, err); } return err; } +static int mlx5e_set_pauseparam(struct net_device *netdev, + struct ethtool_pauseparam *pauseparam) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + + return mlx5e_ethtool_set_pauseparam(priv, pauseparam); +} + int mlx5e_ethtool_get_ts_info(struct mlx5e_priv *priv, struct ethtool_ts_info *info) { -- cgit v1.2.3-55-g7522 From ff9b85de5d5d292fa07869f194fd9e3966f1e720 Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Tue, 6 Nov 2018 19:49:22 +0200 Subject: net/mlx5e: Add some ethtool port control entries to the uplink rep netdev Some of the ethtool entries to control the port should be supported by the uplink rep netdev in switchdev mode, add them. While here, add also the get/set coalesce entries for all reps. Signed-off-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 75 +++++++++++++++++++++++- 1 file changed, 73 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index f557aed54564..ed1158b58798 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -292,6 +292,22 @@ static int mlx5e_rep_set_channels(struct net_device *dev, return 0; } +static int mlx5e_rep_get_coalesce(struct net_device *netdev, + struct ethtool_coalesce *coal) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + + return mlx5e_ethtool_get_coalesce(priv, coal); +} + +static int mlx5e_rep_set_coalesce(struct net_device *netdev, + struct ethtool_coalesce *coal) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + + return mlx5e_ethtool_set_coalesce(priv, coal); +} + static u32 mlx5e_rep_get_rxfh_key_size(struct net_device *netdev) { struct mlx5e_priv *priv = netdev_priv(netdev); @@ -306,7 +322,55 @@ static u32 mlx5e_rep_get_rxfh_indir_size(struct net_device *netdev) return mlx5e_ethtool_get_rxfh_indir_size(priv); } -static const struct ethtool_ops mlx5e_rep_ethtool_ops = { +static void mlx5e_uplink_rep_get_pauseparam(struct net_device *netdev, + struct ethtool_pauseparam *pauseparam) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + + mlx5e_ethtool_get_pauseparam(priv, pauseparam); +} + +static int mlx5e_uplink_rep_set_pauseparam(struct net_device *netdev, + struct ethtool_pauseparam *pauseparam) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + + return mlx5e_ethtool_set_pauseparam(priv, pauseparam); +} + +static int mlx5e_uplink_rep_get_link_ksettings(struct net_device *netdev, + struct ethtool_link_ksettings *link_ksettings) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + + return mlx5e_ethtool_get_link_ksettings(priv, link_ksettings); +} + +static int mlx5e_uplink_rep_set_link_ksettings(struct net_device *netdev, + const struct ethtool_link_ksettings *link_ksettings) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + + return mlx5e_ethtool_set_link_ksettings(priv, link_ksettings); +} + +static const struct ethtool_ops mlx5e_vf_rep_ethtool_ops = { + .get_drvinfo = mlx5e_rep_get_drvinfo, + .get_link = ethtool_op_get_link, + .get_strings = mlx5e_rep_get_strings, + .get_sset_count = mlx5e_rep_get_sset_count, + .get_ethtool_stats = mlx5e_rep_get_ethtool_stats, + .get_ringparam = mlx5e_rep_get_ringparam, + .set_ringparam = mlx5e_rep_set_ringparam, + .get_channels = mlx5e_rep_get_channels, + .set_channels = mlx5e_rep_set_channels, + .get_coalesce = mlx5e_rep_get_coalesce, + .set_coalesce = mlx5e_rep_set_coalesce, + .get_rxfh_key_size = mlx5e_rep_get_rxfh_key_size, + .get_rxfh_indir_size = mlx5e_rep_get_rxfh_indir_size, +}; + +static const struct ethtool_ops mlx5e_uplink_rep_ethtool_ops = { .get_drvinfo = mlx5e_rep_get_drvinfo, .get_link = ethtool_op_get_link, .get_strings = mlx5e_rep_get_strings, @@ -316,8 +380,14 @@ static const struct ethtool_ops mlx5e_rep_ethtool_ops = { .set_ringparam = mlx5e_rep_set_ringparam, .get_channels = mlx5e_rep_get_channels, .set_channels = mlx5e_rep_set_channels, + .get_coalesce = mlx5e_rep_get_coalesce, + .set_coalesce = mlx5e_rep_set_coalesce, + .get_link_ksettings = mlx5e_uplink_rep_get_link_ksettings, + .set_link_ksettings = mlx5e_uplink_rep_set_link_ksettings, .get_rxfh_key_size = mlx5e_rep_get_rxfh_key_size, .get_rxfh_indir_size = mlx5e_rep_get_rxfh_indir_size, + .get_pauseparam = mlx5e_uplink_rep_get_pauseparam, + .set_pauseparam = mlx5e_uplink_rep_set_pauseparam, }; static int mlx5e_attr_get(struct net_device *dev, struct switchdev_attr *attr) @@ -1295,6 +1365,7 @@ static void mlx5e_build_rep_netdev(struct net_device *netdev) netdev->netdev_ops = &mlx5e_netdev_ops_uplink_rep; /* we want a persistent mac for the uplink rep */ mlx5_query_nic_vport_mac_address(mdev, 0, netdev->dev_addr); + netdev->ethtool_ops = &mlx5e_uplink_rep_ethtool_ops; #ifdef CONFIG_MLX5_CORE_EN_DCB if (MLX5_CAP_GEN(mdev, qos)) netdev->dcbnl_ops = &mlx5e_dcbnl_ops; @@ -1302,11 +1373,11 @@ static void mlx5e_build_rep_netdev(struct net_device *netdev) } else { netdev->netdev_ops = &mlx5e_netdev_ops_vf_rep; eth_hw_addr_random(netdev); + netdev->ethtool_ops = &mlx5e_vf_rep_ethtool_ops; } netdev->watchdog_timeo = 15 * HZ; - netdev->ethtool_ops = &mlx5e_rep_ethtool_ops; netdev->switchdev_ops = &mlx5e_rep_switchdev_ops; -- cgit v1.2.3-55-g7522 From 6362a6a04031cacc8ff54b65b84385d9fbab853d Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Tue, 18 Dec 2018 17:15:22 +0100 Subject: drivers: net: ethernet: mellanox: use skb_sec_path helper Will avoid touching this when sp pointer is removed from sk_buff struct. Signed-off-by: Florian Westphal Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c index 128a82b1dbfc..f6717c287ff4 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c @@ -254,11 +254,13 @@ struct sk_buff *mlx5e_ipsec_handle_tx_skb(struct net_device *netdev, struct mlx5e_ipsec_metadata *mdata; struct mlx5e_ipsec_sa_entry *sa_entry; struct xfrm_state *x; + struct sec_path *sp; if (!xo) return skb; - if (unlikely(skb->sp->len != 1)) { + sp = skb_sec_path(skb); + if (unlikely(sp->len != 1)) { atomic64_inc(&priv->ipsec->sw_stats.ipsec_tx_drop_bundle); goto drop; } @@ -372,10 +374,11 @@ struct sk_buff *mlx5e_ipsec_handle_rx_skb(struct net_device *netdev, bool mlx5e_ipsec_feature_check(struct sk_buff *skb, struct net_device *netdev, netdev_features_t features) { + struct sec_path *sp = skb_sec_path(skb); struct xfrm_state *x; - if (skb->sp && skb->sp->len) { - x = skb->sp->xvec[0]; + if (sp && sp->len) { + x = sp->xvec[0]; if (x && x->xso.offload_handle) return true; } -- cgit v1.2.3-55-g7522 From a84e3f533324e40e4a99f50dee2188bf140d8098 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Tue, 18 Dec 2018 17:15:26 +0100 Subject: xfrm: prefer secpath_set over secpath_dup secpath_set is a wrapper for secpath_dup that will not perform an allocation if the secpath attached to the skb has a reference count of one, i.e., it doesn't need to be COW'ed. Also, secpath_dup doesn't attach the secpath to the skb, it leaves this to the caller. Use secpath_set in places that immediately assign the return value to skb. This allows to remove skb->sp without touching these spots again. secpath_dup can eventually be removed in followup patch. Signed-off-by: Florian Westphal Signed-off-by: David S. Miller --- drivers/net/ethernet/intel/ixgbe/ixgbe_ipsec.c | 9 +++++---- drivers/net/ethernet/intel/ixgbevf/ipsec.c | 9 +++++---- drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c | 10 ++++++---- net/xfrm/Kconfig | 1 + net/xfrm/xfrm_output.c | 7 ++----- 5 files changed, 19 insertions(+), 17 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ipsec.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ipsec.c index 8befc7a50f8c..ff85ce5791a3 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_ipsec.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ipsec.c @@ -1161,6 +1161,7 @@ void ixgbe_ipsec_rx(struct ixgbe_ring *rx_ring, struct xfrm_state *xs = NULL; struct ipv6hdr *ip6 = NULL; struct iphdr *ip4 = NULL; + struct sec_path *sp; void *daddr; __be32 spi; u8 *c_hdr; @@ -1200,12 +1201,12 @@ void ixgbe_ipsec_rx(struct ixgbe_ring *rx_ring, if (unlikely(!xs)) return; - skb->sp = secpath_dup(skb->sp); - if (unlikely(!skb->sp)) + sp = secpath_set(skb); + if (unlikely(!sp)) return; - skb->sp->xvec[skb->sp->len++] = xs; - skb->sp->olen++; + sp->xvec[sp->len++] = xs; + sp->olen++; xo = xfrm_offload(skb); xo->flags = CRYPTO_DONE; xo->status = CRYPTO_SUCCESS; diff --git a/drivers/net/ethernet/intel/ixgbevf/ipsec.c b/drivers/net/ethernet/intel/ixgbevf/ipsec.c index 07644e6bf498..5170dd9d8705 100644 --- a/drivers/net/ethernet/intel/ixgbevf/ipsec.c +++ b/drivers/net/ethernet/intel/ixgbevf/ipsec.c @@ -548,6 +548,7 @@ void ixgbevf_ipsec_rx(struct ixgbevf_ring *rx_ring, struct xfrm_state *xs = NULL; struct ipv6hdr *ip6 = NULL; struct iphdr *ip4 = NULL; + struct sec_path *sp; void *daddr; __be32 spi; u8 *c_hdr; @@ -587,12 +588,12 @@ void ixgbevf_ipsec_rx(struct ixgbevf_ring *rx_ring, if (unlikely(!xs)) return; - skb->sp = secpath_dup(skb->sp); - if (unlikely(!skb->sp)) + sp = secpath_set(skb); + if (unlikely(!sp)) return; - skb->sp->xvec[skb->sp->len++] = xs; - skb->sp->olen++; + sp->xvec[sp->len++] = xs; + sp->olen++; xo = xfrm_offload(skb); xo->flags = CRYPTO_DONE; xo->status = CRYPTO_SUCCESS; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c index f6717c287ff4..53608afd39b6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c @@ -307,10 +307,11 @@ mlx5e_ipsec_build_sp(struct net_device *netdev, struct sk_buff *skb, struct mlx5e_priv *priv = netdev_priv(netdev); struct xfrm_offload *xo; struct xfrm_state *xs; + struct sec_path *sp; u32 sa_handle; - skb->sp = secpath_dup(skb->sp); - if (unlikely(!skb->sp)) { + sp = secpath_set(skb); + if (unlikely(!sp)) { atomic64_inc(&priv->ipsec->sw_stats.ipsec_rx_drop_sp_alloc); return NULL; } @@ -322,8 +323,9 @@ mlx5e_ipsec_build_sp(struct net_device *netdev, struct sk_buff *skb, return NULL; } - skb->sp->xvec[skb->sp->len++] = xs; - skb->sp->olen++; + sp = skb_sec_path(skb); + sp->xvec[sp->len++] = xs; + sp->olen++; xo = xfrm_offload(skb); xo->flags = CRYPTO_DONE; diff --git a/net/xfrm/Kconfig b/net/xfrm/Kconfig index 140270a13d54..5d43aaa17027 100644 --- a/net/xfrm/Kconfig +++ b/net/xfrm/Kconfig @@ -5,6 +5,7 @@ config XFRM bool depends on NET select GRO_CELLS + select SKB_EXTENSIONS config XFRM_OFFLOAD bool diff --git a/net/xfrm/xfrm_output.c b/net/xfrm/xfrm_output.c index 4ae87c5ce2e3..757c4d11983b 100644 --- a/net/xfrm/xfrm_output.c +++ b/net/xfrm/xfrm_output.c @@ -218,19 +218,16 @@ int xfrm_output(struct sock *sk, struct sk_buff *skb) if (xfrm_dev_offload_ok(skb, x)) { struct sec_path *sp; - sp = secpath_dup(skb->sp); + sp = secpath_set(skb); if (!sp) { XFRM_INC_STATS(net, LINUX_MIB_XFRMOUTERROR); kfree_skb(skb); return -ENOMEM; } - if (skb->sp) - secpath_put(skb->sp); - skb->sp = sp; skb->encapsulation = 1; sp->olen++; - sp->xvec[skb->sp->len++] = x; + sp->xvec[sp->len++] = x; xfrm_state_hold(x); if (skb_is_gso(skb)) { -- cgit v1.2.3-55-g7522 From 5d8678365c90b9ce1fd2243ff5ea562609f6cec1 Mon Sep 17 00:00:00 2001 From: Miroslav Lichvar Date: Mon, 3 Dec 2018 13:59:41 +0100 Subject: mlx5: update timecounter at least twice per counter overflow The timecounter needs to be updated at least once in half of the cyclecounter interval to prevent timecounter_cyc2time() interpreting a new timestamp as an old value and causing a backward jump. This would be an issue if the timecounter multiplier was so small that the update interval would not be limited by the 64-bit overflow in multiplication. Shorten the calculated interval to make sure the timecounter is updated in time even when the system clock is slowed down by up to 10%, the multiplier is increased by up to 10%, and the scheduled overflow check is late by 15%. Cc: Richard Cochran Cc: Ariel Levkovich Cc: Saeed Mahameed Signed-off-by: Miroslav Lichvar Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c index d27c239e7d6c..53f021df4934 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c @@ -517,14 +517,14 @@ void mlx5_init_clock(struct mlx5_core_dev *mdev) ktime_to_ns(ktime_get_real())); /* Calculate period in seconds to call the overflow watchdog - to make - * sure counter is checked at least once every wrap around. + * sure counter is checked at least twice every wrap around. * The period is calculated as the minimum between max HW cycles count * (The clock source mask) and max amount of cycles that can be * multiplied by clock multiplier where the result doesn't exceed * 64bits. */ overflow_cycles = div64_u64(~0ULL >> 1, clock->cycles.mult); - overflow_cycles = min(overflow_cycles, clock->cycles.mask >> 1); + overflow_cycles = min(overflow_cycles, div_u64(clock->cycles.mask, 3)); ns = cyclecounter_cyc2ns(&clock->cycles, overflow_cycles, frac, &frac); -- cgit v1.2.3-55-g7522 From 4a0475d57ad134429f9bb56068bb738b4afc5992 Mon Sep 17 00:00:00 2001 From: Miroslav Lichvar Date: Mon, 3 Dec 2018 13:59:42 +0100 Subject: mlx5: extend PTP gettime function to read system clock Read the system time right before and immediately after reading the low register of the internal timer. This adds support for the PTP_SYS_OFFSET_EXTENDED ioctl. Cc: Richard Cochran Cc: Saeed Mahameed Signed-off-by: Miroslav Lichvar Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c | 20 ++++++++++++-------- drivers/net/ethernet/mellanox/mlx5/core/main.c | 11 +++++++++-- drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h | 4 +++- 3 files changed, 24 insertions(+), 11 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c index 53f021df4934..ca0ee9916e9e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c @@ -72,7 +72,7 @@ static u64 read_internal_timer(const struct cyclecounter *cc) struct mlx5_core_dev *mdev = container_of(clock, struct mlx5_core_dev, clock); - return mlx5_read_internal_timer(mdev) & cc->mask; + return mlx5_read_internal_timer(mdev, NULL) & cc->mask; } static void mlx5_update_clock_info_page(struct mlx5_core_dev *mdev) @@ -156,15 +156,19 @@ static int mlx5_ptp_settime(struct ptp_clock_info *ptp, return 0; } -static int mlx5_ptp_gettime(struct ptp_clock_info *ptp, struct timespec64 *ts) +static int mlx5_ptp_gettimex(struct ptp_clock_info *ptp, struct timespec64 *ts, + struct ptp_system_timestamp *sts) { struct mlx5_clock *clock = container_of(ptp, struct mlx5_clock, ptp_info); - u64 ns; + struct mlx5_core_dev *mdev = container_of(clock, struct mlx5_core_dev, + clock); unsigned long flags; + u64 cycles, ns; write_seqlock_irqsave(&clock->lock, flags); - ns = timecounter_read(&clock->tc); + cycles = mlx5_read_internal_timer(mdev, sts); + ns = timecounter_cyc2time(&clock->tc, cycles); write_sequnlock_irqrestore(&clock->lock, flags); *ts = ns_to_timespec64(ns); @@ -307,7 +311,7 @@ static int mlx5_perout_configure(struct ptp_clock_info *ptp, ts.tv_sec = rq->perout.start.sec; ts.tv_nsec = rq->perout.start.nsec; ns = timespec64_to_ns(&ts); - cycles_now = mlx5_read_internal_timer(mdev); + cycles_now = mlx5_read_internal_timer(mdev, NULL); write_seqlock_irqsave(&clock->lock, flags); nsec_now = timecounter_cyc2time(&clock->tc, cycles_now); nsec_delta = ns - nsec_now; @@ -384,7 +388,7 @@ static const struct ptp_clock_info mlx5_ptp_clock_info = { .pps = 0, .adjfreq = mlx5_ptp_adjfreq, .adjtime = mlx5_ptp_adjtime, - .gettime64 = mlx5_ptp_gettime, + .gettimex64 = mlx5_ptp_gettimex, .settime64 = mlx5_ptp_settime, .enable = NULL, .verify = NULL, @@ -469,8 +473,8 @@ static int mlx5_pps_event(struct notifier_block *nb, ptp_clock_event(clock->ptp, &ptp_event); break; case PTP_PF_PEROUT: - mlx5_ptp_gettime(&clock->ptp_info, &ts); - cycles_now = mlx5_read_internal_timer(mdev); + mlx5_ptp_gettimex(&clock->ptp_info, &ts, NULL); + cycles_now = mlx5_read_internal_timer(mdev, NULL); ts.tv_sec += 1; ts.tv_nsec = 0; ns = timespec64_to_ns(&ts); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index c23553164e0d..77896c11f6f3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -580,15 +580,22 @@ int mlx5_core_disable_hca(struct mlx5_core_dev *dev, u16 func_id) return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); } -u64 mlx5_read_internal_timer(struct mlx5_core_dev *dev) +u64 mlx5_read_internal_timer(struct mlx5_core_dev *dev, + struct ptp_system_timestamp *sts) { u32 timer_h, timer_h1, timer_l; timer_h = ioread32be(&dev->iseg->internal_timer_h); + ptp_read_system_prets(sts); timer_l = ioread32be(&dev->iseg->internal_timer_l); + ptp_read_system_postts(sts); timer_h1 = ioread32be(&dev->iseg->internal_timer_h); - if (timer_h != timer_h1) /* wrap around */ + if (timer_h != timer_h1) { + /* wrap around */ + ptp_read_system_prets(sts); timer_l = ioread32be(&dev->iseg->internal_timer_l); + ptp_read_system_postts(sts); + } return (u64)timer_l | (u64)timer_h1 << 32; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index 73bf46599ec6..c68dcea5985b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -38,6 +38,7 @@ #include #include #include +#include #include #include @@ -121,7 +122,8 @@ int mlx5_modify_scheduling_element_cmd(struct mlx5_core_dev *dev, u8 hierarchy, int mlx5_destroy_scheduling_element_cmd(struct mlx5_core_dev *dev, u8 hierarchy, u32 element_id); int mlx5_wait_for_vf_pages(struct mlx5_core_dev *dev); -u64 mlx5_read_internal_timer(struct mlx5_core_dev *dev); +u64 mlx5_read_internal_timer(struct mlx5_core_dev *dev, + struct ptp_system_timestamp *sts); void mlx5_cmd_trigger_completions(struct mlx5_core_dev *dev); int mlx5_cq_debugfs_init(struct mlx5_core_dev *dev); -- cgit v1.2.3-55-g7522 From 4c8fb2986d44c5a75949a88ac61deec0ed50a542 Mon Sep 17 00:00:00 2001 From: Gavi Teitz Date: Wed, 12 Dec 2018 21:23:18 +0200 Subject: net/mlx5e: Increase VF representors' SQ size to 128 The default size for the VF representors' SQ was too small to handle high packet rates. Doubling the size from 64 to 128 drastically improves the packet rate under stress (by about 50%), whereas increasing the size beyond 128 has not shown to make any further difference. The impact of the SQ size was measured with UDP traffic, in the following topology: TG <-> PF <-> TC forwarding <-> VF representor <-> VF in VM over a single core processing bi-directional traffic, with the following results: SQ size of 64: SQ size of 128: Packet rate for 64B UDP packets: 860 [Kpps] 1280 [Kpps] Packet rate for 114B VxLan encapsulated UDP packets: 320 [Kpps] 500 [Kpps] Signed-off-by: Gavi Teitz Reviewed-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index ed1158b58798..18d0f4f7537b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -45,8 +45,8 @@ #include "en/tc_tun.h" #include "fs_core.h" -#define MLX5E_REP_PARAMS_LOG_SQ_SIZE \ - max(0x6, MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE) +#define MLX5E_REP_PARAMS_DEF_LOG_SQ_SIZE \ + max(0x7, MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE) static const char mlx5e_rep_driver_name[] = "mlx5e_rep"; @@ -1336,7 +1336,7 @@ static void mlx5e_build_rep_params(struct net_device *netdev) if (rep->vport == FDB_UPLINK_VPORT) params->log_sq_size = MLX5E_PARAMS_DEFAULT_LOG_SQ_SIZE; else - params->log_sq_size = MLX5E_REP_PARAMS_LOG_SQ_SIZE; + params->log_sq_size = MLX5E_REP_PARAMS_DEF_LOG_SQ_SIZE;; /* RQ */ mlx5e_build_rq_params(mdev, params); -- cgit v1.2.3-55-g7522 From 1ee4457c5c2a6172893eb1f464860c753c9b4e96 Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Tue, 18 Dec 2018 13:32:46 +0200 Subject: net/mlx5e: Allow vlans on e-switch uplink reps There are cases (e.g tunneling with vlan on underlay and potentially more) where this makes sense, so allow that. Signed-off-by: Or Gerlitz Reviewed-by: Eli Britstein Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index 18d0f4f7537b..91c3eb85f32e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1381,7 +1381,7 @@ static void mlx5e_build_rep_netdev(struct net_device *netdev) netdev->switchdev_ops = &mlx5e_rep_switchdev_ops; - netdev->features |= NETIF_F_VLAN_CHALLENGED | NETIF_F_HW_TC | NETIF_F_NETNS_LOCAL; + netdev->features |= NETIF_F_HW_TC | NETIF_F_NETNS_LOCAL; netdev->hw_features |= NETIF_F_HW_TC; netdev->hw_features |= NETIF_F_SG; @@ -1392,6 +1392,9 @@ static void mlx5e_build_rep_netdev(struct net_device *netdev) netdev->hw_features |= NETIF_F_TSO6; netdev->hw_features |= NETIF_F_RXCSUM; + if (rep->vport != FDB_UPLINK_VPORT) + netdev->features |= NETIF_F_VLAN_CHALLENGED; + netdev->features |= netdev->hw_features; } -- cgit v1.2.3-55-g7522 From a0646c88ed3ac527b71cf365b03ff4fc729d35b6 Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Wed, 19 Dec 2018 09:24:58 +0200 Subject: net/mlx5e: Fail attempt to offload e-switch TC flows with egress upper devices We use the switchdev parent HW id helper to identify if the mirred device shares the same ASIC/port with the ingress device. This can get us wrong in the presence of upper devices such as vlan or bridge set over the HW devices (VF or uplink representors), b/c the switchdev ID is retrieved recursively. To fail offload attempts in such cases, we condition the check on the egress device to have not only the same switchdev ID but also the relevant mlx5 netdev ops. Fixes: 03a9d11e6eeb ('net/mlx5e: Add TC drop and mirred/redirect action parsing for SRIOV offloads') Signed-off-by: Eli Britstein Reviewed-by: Roi Dayan Acked-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 9 +++++++++ drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | 3 +++ drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 3 +++ 3 files changed, 15 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index 91c3eb85f32e..f414f19c1159 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1316,6 +1316,15 @@ static const struct net_device_ops mlx5e_netdev_ops_uplink_rep = { .ndo_get_vf_stats = mlx5e_get_vf_stats, }; +bool mlx5e_eswitch_rep(struct net_device *netdev) +{ + if (netdev->netdev_ops == &mlx5e_netdev_ops_vf_rep || + netdev->netdev_ops == &mlx5e_netdev_ops_uplink_rep) + return true; + + return false; +} + static void mlx5e_build_rep_params(struct net_device *netdev) { struct mlx5e_priv *priv = netdev_priv(netdev); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h index 5645d3cef1bb..edd722824697 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h @@ -176,6 +176,9 @@ void mlx5e_rep_encap_entry_detach(struct mlx5e_priv *priv, struct mlx5e_encap_entry *e); void mlx5e_rep_queue_neigh_stats_work(struct mlx5e_priv *priv); + +bool mlx5e_eswitch_rep(struct net_device *netdev); + #else /* CONFIG_MLX5_ESWITCH */ static inline bool mlx5e_is_uplink_rep(struct mlx5e_priv *priv) { return false; } static inline int mlx5e_add_sqs_fwd_rules(struct mlx5e_priv *priv) { return 0; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index c1a9120412b8..9ba8ade3be47 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -2583,6 +2583,9 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, uplink_upper == out_dev) out_dev = uplink_dev; + if (!mlx5e_eswitch_rep(out_dev)) + return -EOPNOTSUPP; + out_priv = netdev_priv(out_dev); rpriv = out_priv->ppriv; attr->dests[attr->out_count].rep = rpriv->rep; -- cgit v1.2.3-55-g7522 From 442e1228cbaccee8a9365624608544460155c5d1 Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Tue, 18 Dec 2018 09:46:00 +0200 Subject: net/mlx5e: Tunnel routing output devs helper function For tunnel we determine the output devs for IPv4/6 cases, in two separate functions, with a duplicated code. Move that code from IPv4/6 functions to a helper function, with no functional change. Signed-off-by: Eli Britstein Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/en/tc_tun.c | 70 +++++++++++----------- 1 file changed, 34 insertions(+), 36 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') 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 b92f8b3ff6b2..c90a0f9cb6e1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c @@ -6,6 +6,34 @@ #include "lib/vxlan.h" #include "en/tc_tun.h" +static int get_route_and_out_devs(struct mlx5e_priv *priv, + struct net_device *dev, + struct net_device **route_dev, + struct net_device **out_dev) +{ + struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; + struct net_device *uplink_dev, *uplink_upper; + bool dst_is_lag_dev; + + uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH); + uplink_upper = netdev_master_upper_dev_get(uplink_dev); + dst_is_lag_dev = (uplink_upper && + netif_is_lag_master(uplink_upper) && + dev == uplink_upper && + mlx5_lag_is_sriov(priv->mdev)); + + /* if the egress device isn't on the same HW e-switch or + * it's a LAG device, use the uplink + */ + if (!switchdev_port_same_parent_id(priv->netdev, dev) || + dst_is_lag_dev) + *out_dev = uplink_dev; + else + *out_dev = dev; + + return 0; +} + static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv, struct net_device *mirred_dev, struct net_device **out_dev, @@ -13,9 +41,6 @@ static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv, struct neighbour **out_n, u8 *out_ttl) { - struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; - struct net_device *uplink_dev, *uplink_upper; - bool dst_is_lag_dev; struct rtable *rt; struct neighbour *n = NULL; @@ -30,21 +55,9 @@ static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv, return -EOPNOTSUPP; #endif - uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH); - uplink_upper = netdev_master_upper_dev_get(uplink_dev); - dst_is_lag_dev = (uplink_upper && - netif_is_lag_master(uplink_upper) && - rt->dst.dev == uplink_upper && - mlx5_lag_is_sriov(priv->mdev)); - - /* if the egress device isn't on the same HW e-switch or - * it's a LAG device, use the uplink - */ - if (!switchdev_port_same_parent_id(priv->netdev, rt->dst.dev) || - dst_is_lag_dev) - *out_dev = uplink_dev; - else - *out_dev = rt->dst.dev; + ret = get_route_and_out_devs(priv, rt->dst.dev, NULL, out_dev); + if (ret < 0) + return ret; if (!(*out_ttl)) *out_ttl = ip4_dst_hoplimit(&rt->dst); @@ -76,9 +89,6 @@ static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv, struct dst_entry *dst; #if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6) - struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; - struct net_device *uplink_dev, *uplink_upper; - bool dst_is_lag_dev; int ret; ret = ipv6_stub->ipv6_dst_lookup(dev_net(mirred_dev), NULL, &dst, @@ -89,21 +99,9 @@ static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv, if (!(*out_ttl)) *out_ttl = ip6_dst_hoplimit(dst); - uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH); - uplink_upper = netdev_master_upper_dev_get(uplink_dev); - dst_is_lag_dev = (uplink_upper && - netif_is_lag_master(uplink_upper) && - dst->dev == uplink_upper && - mlx5_lag_is_sriov(priv->mdev)); - - /* if the egress device isn't on the same HW e-switch or - * it's a LAG device, use the uplink - */ - if (!switchdev_port_same_parent_id(priv->netdev, dst->dev) || - dst_is_lag_dev) - *out_dev = uplink_dev; - else - *out_dev = dst->dev; + ret = get_route_and_out_devs(priv, dst->dev, NULL, out_dev); + if (ret < 0) + return ret; #else return -EOPNOTSUPP; #endif -- cgit v1.2.3-55-g7522 From b168cff0b9b216e68215cdb28db999f542482a56 Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Wed, 19 Dec 2018 09:29:10 +0200 Subject: net/mlx5e: Fail attempt to offload e-switch TC encap flows with vlan on underlay Currently we don't support nor fail attempts to offload encap flows routed to vlan device on the underlay network. We wrongly consider a vlan underlay device to be on the same e-switch b/c the switchdev ID is retrieved recursively. Add explicit check for that and fail such attempts. Also align to a more strict check for the ingress and the underlay devices to practically be on the same eswitch. Fixes: ce99f6b97fcd ('net/mlx5e: Support SRIOV TC encapsulation offloads for IPv6 tunnels') Fixes: 3e621b19b0bb ('net/mlx5e: Support TC encapsulation offloads with upper devices') Signed-off-by: Eli Britstein Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') 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 c90a0f9cb6e1..0d24e9648c21 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c @@ -15,6 +15,10 @@ static int get_route_and_out_devs(struct mlx5e_priv *priv, struct net_device *uplink_dev, *uplink_upper; bool dst_is_lag_dev; + /* we currently don't offload vlan on underlay */ + if (is_vlan_dev(dev)) + return -EOPNOTSUPP; + uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH); uplink_upper = netdev_master_upper_dev_get(uplink_dev); dst_is_lag_dev = (uplink_upper && @@ -28,6 +32,8 @@ static int get_route_and_out_devs(struct mlx5e_priv *priv, if (!switchdev_port_same_parent_id(priv->netdev, dev) || dst_is_lag_dev) *out_dev = uplink_dev; + else if (!mlx5e_eswitch_rep(dev)) + return -EOPNOTSUPP; else *out_dev = dev; -- cgit v1.2.3-55-g7522 From 05ada1adb65acd533f978651e7001e3568e7b98a Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Sun, 9 Dec 2018 09:17:18 +0200 Subject: net/mlx5e: Tunnel encap ETH header helper function In tunnel encap we prepare the encap header for IPv4/6 cases, in two separate functions. For ETH header generation the code is almost duplicated. Move the ETH header generation code from IPv4/6 functions to a helper function, with no functional change. Signed-off-by: Eli Britstein Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/en/tc_tun.c | 29 +++++++++++++--------- 1 file changed, 17 insertions(+), 12 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') 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 0d24e9648c21..f90c67c9add0 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c @@ -180,6 +180,19 @@ static int mlx5e_gen_ip_tunnel_header(char buf[], __u8 *ip_proto, return err; } +static char *gen_eth_tnl_hdr(char *buf, struct net_device *dev, + struct mlx5e_encap_entry *e, + u16 proto) +{ + struct ethhdr *eth = (struct ethhdr *)buf; + + ether_addr_copy(eth->h_dest, e->h_dest); + ether_addr_copy(eth->h_source, dev->dev_addr); + eth->h_proto = htons(proto); + + return (char *)eth + ETH_HLEN; +} + int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv, struct net_device *mirred_dev, struct mlx5e_encap_entry *e) @@ -193,7 +206,6 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv, struct neighbour *n = NULL; struct flowi4 fl4 = {}; char *encap_header; - struct ethhdr *eth; u8 nud_state, ttl; struct iphdr *ip; int err; @@ -242,13 +254,10 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv, read_unlock_bh(&n->lock); /* add ethernet header */ - eth = (struct ethhdr *)encap_header; - ether_addr_copy(eth->h_dest, e->h_dest); - ether_addr_copy(eth->h_source, out_dev->dev_addr); - eth->h_proto = htons(ETH_P_IP); + ip = (struct iphdr *)gen_eth_tnl_hdr(encap_header, out_dev, e, + ETH_P_IP); /* add ip header */ - ip = (struct iphdr *)((char *)eth + sizeof(struct ethhdr)); ip->tos = tun_key->tos; ip->version = 0x4; ip->ihl = 0x5; @@ -308,7 +317,6 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv, struct flowi6 fl6 = {}; struct ipv6hdr *ip6h; char *encap_header; - struct ethhdr *eth; u8 nud_state, ttl; int err; @@ -356,13 +364,10 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv, read_unlock_bh(&n->lock); /* add ethernet header */ - eth = (struct ethhdr *)encap_header; - ether_addr_copy(eth->h_dest, e->h_dest); - ether_addr_copy(eth->h_source, out_dev->dev_addr); - eth->h_proto = htons(ETH_P_IPV6); + ip6h = (struct ipv6hdr *)gen_eth_tnl_hdr(encap_header, out_dev, e, + ETH_P_IPV6); /* add ip header */ - ip6h = (struct ipv6hdr *)((char *)eth + sizeof(struct ethhdr)); ip6_flow_hdr(ip6h, tun_key->tos, 0); /* the HW fills up ipv6 payload len */ ip6h->hop_limit = ttl; -- cgit v1.2.3-55-g7522 From c7bcb277bda46d75819fb2d475fa817c10dc209b Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Mon, 3 Dec 2018 17:09:54 +0200 Subject: net/mlx5e: Re-order route and encap header memory allocation Change the order to first route IPv4/6 and return if error. Only after successful route continue to allocate an encap header, with no functional change. Signed-off-by: Eli Britstein Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/en/tc_tun.c | 62 ++++++++++++---------- 1 file changed, 34 insertions(+), 28 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') 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 f90c67c9add0..bcf08f5abc2f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c @@ -198,28 +198,16 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv, struct mlx5e_encap_entry *e) { int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size); - int ipv4_encap_size = ETH_HLEN + - sizeof(struct iphdr) + - e->tunnel_hlen; struct ip_tunnel_key *tun_key = &e->tun_info.key; struct net_device *out_dev; struct neighbour *n = NULL; struct flowi4 fl4 = {}; + int ipv4_encap_size; char *encap_header; u8 nud_state, ttl; struct iphdr *ip; int err; - if (max_encap_size < ipv4_encap_size) { - mlx5_core_warn(priv->mdev, "encap size %d too big, max supported is %d\n", - ipv4_encap_size, max_encap_size); - return -EOPNOTSUPP; - } - - encap_header = kzalloc(ipv4_encap_size, GFP_KERNEL); - if (!encap_header) - return -ENOMEM; - /* add the IP fields */ fl4.flowi4_tos = tun_key->tos; fl4.daddr = tun_key->u.ipv4.dst; @@ -229,7 +217,22 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv, err = mlx5e_route_lookup_ipv4(priv, mirred_dev, &out_dev, &fl4, &n, &ttl); if (err) - goto free_encap; + return err; + + ipv4_encap_size = + ETH_HLEN + + sizeof(struct iphdr) + + e->tunnel_hlen; + + if (max_encap_size < ipv4_encap_size) { + mlx5_core_warn(priv->mdev, "encap size %d too big, max supported is %d\n", + ipv4_encap_size, max_encap_size); + return -EOPNOTSUPP; + } + + encap_header = kzalloc(ipv4_encap_size, GFP_KERNEL); + if (!encap_header) + return -ENOMEM; /* used by mlx5e_detach_encap to lookup a neigh hash table * entry in the neigh hash table when a user deletes a rule @@ -308,28 +311,16 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv, struct mlx5e_encap_entry *e) { int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size); - int ipv6_encap_size = ETH_HLEN + - sizeof(struct ipv6hdr) + - e->tunnel_hlen; struct ip_tunnel_key *tun_key = &e->tun_info.key; struct net_device *out_dev; struct neighbour *n = NULL; struct flowi6 fl6 = {}; struct ipv6hdr *ip6h; + int ipv6_encap_size; char *encap_header; u8 nud_state, ttl; int err; - if (max_encap_size < ipv6_encap_size) { - mlx5_core_warn(priv->mdev, "encap size %d too big, max supported is %d\n", - ipv6_encap_size, max_encap_size); - return -EOPNOTSUPP; - } - - encap_header = kzalloc(ipv6_encap_size, GFP_KERNEL); - if (!encap_header) - return -ENOMEM; - ttl = tun_key->ttl; fl6.flowlabel = ip6_make_flowinfo(RT_TOS(tun_key->tos), tun_key->label); @@ -339,7 +330,22 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv, err = mlx5e_route_lookup_ipv6(priv, mirred_dev, &out_dev, &fl6, &n, &ttl); if (err) - goto free_encap; + return err; + + ipv6_encap_size = + ETH_HLEN + + sizeof(struct ipv6hdr) + + e->tunnel_hlen; + + if (max_encap_size < ipv6_encap_size) { + mlx5_core_warn(priv->mdev, "encap size %d too big, max supported is %d\n", + ipv6_encap_size, max_encap_size); + return -EOPNOTSUPP; + } + + encap_header = kzalloc(ipv6_encap_size, GFP_KERNEL); + if (!encap_header) + return -ENOMEM; /* used by mlx5e_detach_encap to lookup a neigh hash table * entry in the neigh hash table when a user deletes a rule -- cgit v1.2.3-55-g7522 From aa331450b81f071bc8d7b68d5baab1b78f93598a Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Sun, 9 Dec 2018 09:17:18 +0200 Subject: net/mlx5e: Support VLAN encap ETH header generation Support generation of native or tagged Ethernet header for encap header, depending on provided net device. Signed-off-by: Eli Britstein Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') 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 bcf08f5abc2f..a6990744730a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c @@ -185,12 +185,23 @@ static char *gen_eth_tnl_hdr(char *buf, struct net_device *dev, u16 proto) { struct ethhdr *eth = (struct ethhdr *)buf; + char *ip; ether_addr_copy(eth->h_dest, e->h_dest); ether_addr_copy(eth->h_source, dev->dev_addr); - eth->h_proto = htons(proto); + if (is_vlan_dev(dev)) { + struct vlan_hdr *vlan = (struct vlan_hdr *) + ((char *)eth + ETH_HLEN); + ip = (char *)vlan + VLAN_HLEN; + eth->h_proto = vlan_dev_vlan_proto(dev); + vlan->h_vlan_TCI = htons(vlan_dev_vlan_id(dev)); + vlan->h_vlan_encapsulated_proto = htons(proto); + } else { + eth->h_proto = htons(proto); + ip = (char *)eth + ETH_HLEN; + } - return (char *)eth + ETH_HLEN; + return ip; } int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv, -- cgit v1.2.3-55-g7522 From e32ee6c78efa6a32bff782bbe7a9970b018996ca Mon Sep 17 00:00:00 2001 From: Eli Britstein Date: Mon, 3 Dec 2018 17:09:54 +0200 Subject: net/mlx5e: Support tunnel encap over tagged Ethernet Generate encap header depending on the routed device to support native/tagged Ethernet header. Signed-off-by: Eli Britstein Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/en/tc_tun.c | 44 ++++++++++++---------- 1 file changed, 24 insertions(+), 20 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') 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 a6990744730a..046948ead152 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c @@ -15,10 +15,6 @@ static int get_route_and_out_devs(struct mlx5e_priv *priv, struct net_device *uplink_dev, *uplink_upper; bool dst_is_lag_dev; - /* we currently don't offload vlan on underlay */ - if (is_vlan_dev(dev)) - return -EOPNOTSUPP; - uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH); uplink_upper = netdev_master_upper_dev_get(uplink_dev); dst_is_lag_dev = (uplink_upper && @@ -30,12 +26,18 @@ static int get_route_and_out_devs(struct mlx5e_priv *priv, * it's a LAG device, use the uplink */ if (!switchdev_port_same_parent_id(priv->netdev, dev) || - dst_is_lag_dev) - *out_dev = uplink_dev; - else if (!mlx5e_eswitch_rep(dev)) - return -EOPNOTSUPP; - else - *out_dev = dev; + dst_is_lag_dev) { + *route_dev = uplink_dev; + *out_dev = *route_dev; + } else { + *route_dev = dev; + if (is_vlan_dev(*route_dev)) + *out_dev = uplink_dev; + else if (mlx5e_eswitch_rep(dev)) + *out_dev = *route_dev; + else + return -EOPNOTSUPP; + } return 0; } @@ -43,6 +45,7 @@ static int get_route_and_out_devs(struct mlx5e_priv *priv, static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv, struct net_device *mirred_dev, struct net_device **out_dev, + struct net_device **route_dev, struct flowi4 *fl4, struct neighbour **out_n, u8 *out_ttl) @@ -61,7 +64,7 @@ static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv, return -EOPNOTSUPP; #endif - ret = get_route_and_out_devs(priv, rt->dst.dev, NULL, out_dev); + ret = get_route_and_out_devs(priv, rt->dst.dev, route_dev, out_dev); if (ret < 0) return ret; @@ -87,6 +90,7 @@ static const char *mlx5e_netdev_kind(struct net_device *dev) static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv, struct net_device *mirred_dev, struct net_device **out_dev, + struct net_device **route_dev, struct flowi6 *fl6, struct neighbour **out_n, u8 *out_ttl) @@ -105,7 +109,7 @@ static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv, if (!(*out_ttl)) *out_ttl = ip6_dst_hoplimit(dst); - ret = get_route_and_out_devs(priv, dst->dev, NULL, out_dev); + ret = get_route_and_out_devs(priv, dst->dev, route_dev, out_dev); if (ret < 0) return ret; #else @@ -210,7 +214,7 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv, { int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size); struct ip_tunnel_key *tun_key = &e->tun_info.key; - struct net_device *out_dev; + struct net_device *out_dev, *route_dev; struct neighbour *n = NULL; struct flowi4 fl4 = {}; int ipv4_encap_size; @@ -225,13 +229,13 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv, fl4.saddr = tun_key->u.ipv4.src; ttl = tun_key->ttl; - err = mlx5e_route_lookup_ipv4(priv, mirred_dev, &out_dev, + err = mlx5e_route_lookup_ipv4(priv, mirred_dev, &out_dev, &route_dev, &fl4, &n, &ttl); if (err) return err; ipv4_encap_size = - ETH_HLEN + + (is_vlan_dev(route_dev) ? VLAN_ETH_HLEN : ETH_HLEN) + sizeof(struct iphdr) + e->tunnel_hlen; @@ -268,7 +272,7 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv, read_unlock_bh(&n->lock); /* add ethernet header */ - ip = (struct iphdr *)gen_eth_tnl_hdr(encap_header, out_dev, e, + ip = (struct iphdr *)gen_eth_tnl_hdr(encap_header, route_dev, e, ETH_P_IP); /* add ip header */ @@ -323,7 +327,7 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv, { int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size); struct ip_tunnel_key *tun_key = &e->tun_info.key; - struct net_device *out_dev; + struct net_device *out_dev, *route_dev; struct neighbour *n = NULL; struct flowi6 fl6 = {}; struct ipv6hdr *ip6h; @@ -338,13 +342,13 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv, fl6.daddr = tun_key->u.ipv6.dst; fl6.saddr = tun_key->u.ipv6.src; - err = mlx5e_route_lookup_ipv6(priv, mirred_dev, &out_dev, + err = mlx5e_route_lookup_ipv6(priv, mirred_dev, &out_dev, &route_dev, &fl6, &n, &ttl); if (err) return err; ipv6_encap_size = - ETH_HLEN + + (is_vlan_dev(route_dev) ? VLAN_ETH_HLEN : ETH_HLEN) + sizeof(struct ipv6hdr) + e->tunnel_hlen; @@ -381,7 +385,7 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv, read_unlock_bh(&n->lock); /* add ethernet header */ - ip6h = (struct ipv6hdr *)gen_eth_tnl_hdr(encap_header, out_dev, e, + ip6h = (struct ipv6hdr *)gen_eth_tnl_hdr(encap_header, route_dev, e, ETH_P_IPV6); /* add ip header */ -- cgit v1.2.3-55-g7522 From 0a5b589111c395f2b18c2a508e241207daaa8cca Mon Sep 17 00:00:00 2001 From: Aviv Heller Date: Thu, 20 Dec 2018 01:37:17 +0200 Subject: net/mlx5: Fix query_nic_sys_image_guid() error during init vport system image guid should be queried using vport nic API for Ethernet ports, and vport hca API for Infiniband ports. Fixes: fadd59fc50d0 ("net/mlx5: Introduce inter-device communication mechanism") Signed-off-by: Aviv Heller Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/vport.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/vport.c b/drivers/net/ethernet/mellanox/mlx5/core/vport.c index cfbea66b4879..9b150ce9d315 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/vport.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/vport.c @@ -1204,9 +1204,19 @@ EXPORT_SYMBOL_GPL(mlx5_nic_vport_unaffiliate_multiport); u64 mlx5_query_nic_system_image_guid(struct mlx5_core_dev *mdev) { - if (!mdev->sys_image_guid) - mlx5_query_nic_vport_system_image_guid(mdev, &mdev->sys_image_guid); + int port_type_cap = MLX5_CAP_GEN(mdev, port_type); + u64 tmp = 0; - return mdev->sys_image_guid; + if (mdev->sys_image_guid) + return mdev->sys_image_guid; + + if (port_type_cap == MLX5_CAP_PORT_TYPE_ETH) + mlx5_query_nic_vport_system_image_guid(mdev, &tmp); + else + mlx5_query_hca_vport_system_image_guid(mdev, &tmp); + + mdev->sys_image_guid = tmp; + + return tmp; } EXPORT_SYMBOL_GPL(mlx5_query_nic_system_image_guid); -- cgit v1.2.3-55-g7522 From a64917446eafc7212e962561622d697fce04e9a6 Mon Sep 17 00:00:00 2001 From: Aviv Heller Date: Tue, 18 Dec 2018 19:03:27 +0200 Subject: net/mlx5: Fix LAG requirement when CONFIG_MLX5_ESWITCH is off If CONFIG_MLX5_ESWITCH is not defined, test for SR-IOV being disabled, instead of calling e-switch LAG prereq routine. Since LAG with SRIOV is allowed only when switchdev mode is on. Fixes: eff849b2c669 ("net/mlx5: Allow/disallow LAG according to pre-req only") Signed-off-by: Aviv Heller Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/lag.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c index 5187dc7a72a3..3a6baed722d8 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c @@ -291,12 +291,15 @@ static int mlx5_deactivate_lag(struct mlx5_lag *ldev) static bool mlx5_lag_check_prereq(struct mlx5_lag *ldev) { - if (ldev->pf[0].dev && - ldev->pf[1].dev && - mlx5_esw_lag_prereq(ldev->pf[0].dev, ldev->pf[1].dev)) - return true; - else + if (!ldev->pf[0].dev || !ldev->pf[1].dev) return false; + +#ifdef CONFIG_MLX5_ESWITCH + return mlx5_esw_lag_prereq(ldev->pf[0].dev, ldev->pf[1].dev); +#else + return (!mlx5_sriov_is_enabled(ldev->pf[0].dev) && + !mlx5_sriov_is_enabled(ldev->pf[1].dev)); +#endif } static void mlx5_lag_add_ib_devices(struct mlx5_lag *ldev) -- cgit v1.2.3-55-g7522 From e05b8d4fc30a68540e5c0eef3e2bacd1f27b8172 Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Thu, 15 Nov 2018 11:48:41 +0200 Subject: net/mlx5e: TX, Print opcode in error CQE warning The opcode indicates about the error reason. Printing it helps in debug. Signed-off-by: Tariq Toukan Reviewed-by: Eran Ben Elisha Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tx.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c index f201965b888f..598ad7e4d5c9 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c @@ -459,9 +459,10 @@ static void mlx5e_dump_error_cqe(struct mlx5e_txqsq *sq, u32 ci = mlx5_cqwq_get_ci(&sq->cq.wq); netdev_err(sq->channel->netdev, - "Error cqe on cqn 0x%x, ci 0x%x, sqn 0x%x, syndrome 0x%x, vendor syndrome 0x%x\n", - sq->cq.mcq.cqn, ci, sq->sqn, err_cqe->syndrome, - err_cqe->vendor_err_synd); + "Error cqe on cqn 0x%x, ci 0x%x, sqn 0x%x, opcode 0x%x, syndrome 0x%x, vendor syndrome 0x%x\n", + sq->cq.mcq.cqn, ci, sq->sqn, + get_cqe_opcode((struct mlx5_cqe64 *)err_cqe), + err_cqe->syndrome, err_cqe->vendor_err_synd); mlx5_dump_err_cqe(sq->cq.mdev, err_cqe); } -- cgit v1.2.3-55-g7522 From 4fb2f516186ea25dea326f97880431e14a5b9e9d Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Wed, 21 Nov 2018 13:56:17 +0200 Subject: net/mlx5e: XDP, Precede XDP-related operations in RQ poll by a loaded program check At the end of the RQ polling loop, some XDP-related operations might be required. Before checking them one by one, check if an XDP program is even loaded. Combine all the checks and operations in a single function in xdp files. This saves unnecessary checks for non-XDP flows. Signed-off-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c | 15 +++++++++++++++ drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h | 2 +- drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 12 ++---------- 3 files changed, 18 insertions(+), 11 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c index ad6d471d00dd..cf22ea529c53 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c @@ -305,3 +305,18 @@ int mlx5e_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames, return n - drops; } + +void mlx5e_xdp_rx_poll_complete(struct mlx5e_rq *rq) +{ + struct mlx5e_xdpsq *xdpsq = &rq->xdpsq; + + if (xdpsq->doorbell) { + mlx5e_xmit_xdp_doorbell(xdpsq); + xdpsq->doorbell = false; + } + + if (xdpsq->redirect_flush) { + xdp_do_flush_map(); + xdpsq->redirect_flush = false; + } +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h index 6dfab045925f..eecc0392fcff 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h @@ -44,7 +44,7 @@ bool mlx5e_xdp_handle(struct mlx5e_rq *rq, struct mlx5e_dma_info *di, void *va, u16 *rx_headroom, u32 *len); bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq); void mlx5e_free_xdpsq_descs(struct mlx5e_xdpsq *sq); - +void mlx5e_xdp_rx_poll_complete(struct mlx5e_rq *rq); bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi); int mlx5e_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames, u32 flags); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c index cdce30a95c60..1d0bb5ff8c26 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c @@ -1190,7 +1190,6 @@ mpwrq_cqe_out: int mlx5e_poll_rx_cq(struct mlx5e_cq *cq, int budget) { struct mlx5e_rq *rq = container_of(cq, struct mlx5e_rq, cq); - struct mlx5e_xdpsq *xdpsq = &rq->xdpsq; struct mlx5_cqe64 *cqe; int work_done = 0; @@ -1221,15 +1220,8 @@ int mlx5e_poll_rx_cq(struct mlx5e_cq *cq, int budget) } while ((++work_done < budget) && (cqe = mlx5_cqwq_get_cqe(&cq->wq))); out: - if (xdpsq->doorbell) { - mlx5e_xmit_xdp_doorbell(xdpsq); - xdpsq->doorbell = false; - } - - if (xdpsq->redirect_flush) { - xdp_do_flush_map(); - xdpsq->redirect_flush = false; - } + if (rq->xdp_prog) + mlx5e_xdp_rx_poll_complete(rq); mlx5_cqwq_update_db_record(&cq->wq); -- cgit v1.2.3-55-g7522 From feb2ff9d74f76a174e284255b7e537b6c090c13c Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Wed, 21 Nov 2018 14:04:45 +0200 Subject: net/mlx5e: XDP, Change the XDP SQ redirect indication Do not maintain an SQ state bit to indicate whether an XDP SQ serves redirect operations. Instead, rely on the fact that such an XDP SQ doesn't reside in an RQ instance, while the others do. This info is not known to the XDP SQ functions themselves, and they rely on their callers to distinguish between the cases. Signed-off-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 1 - drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c | 14 ++++---------- drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h | 4 ++-- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 12 +++++------- drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c | 13 +++++++------ 5 files changed, 18 insertions(+), 26 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 72cc204f9a34..15cbf0c6e309 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -344,7 +344,6 @@ enum { MLX5E_SQ_STATE_IPSEC, MLX5E_SQ_STATE_AM, MLX5E_SQ_STATE_TLS, - MLX5E_SQ_STATE_REDIRECT, }; struct mlx5e_sq_wqe_info { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c index cf22ea529c53..4961b8fea111 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c @@ -164,11 +164,10 @@ bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi) return true; } -bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq) +bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq, struct mlx5e_rq *rq) { struct mlx5e_xdpsq *sq; struct mlx5_cqe64 *cqe; - struct mlx5e_rq *rq; bool is_redirect; u16 sqcc; int i; @@ -182,8 +181,7 @@ bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq) if (!cqe) return false; - is_redirect = test_bit(MLX5E_SQ_STATE_REDIRECT, &sq->state); - rq = container_of(sq, struct mlx5e_rq, xdpsq); + is_redirect = !rq; /* sq->cc must be updated only after mlx5_cqwq_update_db_record(), * otherwise a cq overrun may occur @@ -228,13 +226,9 @@ bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq) return (i == MLX5E_TX_CQ_POLL_BUDGET); } -void mlx5e_free_xdpsq_descs(struct mlx5e_xdpsq *sq) +void mlx5e_free_xdpsq_descs(struct mlx5e_xdpsq *sq, struct mlx5e_rq *rq) { - struct mlx5e_rq *rq; - bool is_redirect; - - is_redirect = test_bit(MLX5E_SQ_STATE_REDIRECT, &sq->state); - rq = is_redirect ? NULL : container_of(sq, struct mlx5e_rq, xdpsq); + bool is_redirect = !rq; while (sq->cc != sq->pc) { u16 ci = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->cc); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h index eecc0392fcff..37fcb17e7f27 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h @@ -42,8 +42,8 @@ bool mlx5e_xdp_handle(struct mlx5e_rq *rq, struct mlx5e_dma_info *di, void *va, u16 *rx_headroom, u32 *len); -bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq); -void mlx5e_free_xdpsq_descs(struct mlx5e_xdpsq *sq); +bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq, struct mlx5e_rq *rq); +void mlx5e_free_xdpsq_descs(struct mlx5e_xdpsq *sq, struct mlx5e_rq *rq); void mlx5e_xdp_rx_poll_complete(struct mlx5e_rq *rq); bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi); int mlx5e_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index bc791404f2e7..8a19f451fb7a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -1577,8 +1577,6 @@ static int mlx5e_open_xdpsq(struct mlx5e_channel *c, csp.cqn = sq->cq.mcq.cqn; csp.wq_ctrl = &sq->wq_ctrl; csp.min_inline_mode = sq->min_inline_mode; - if (is_redirect) - set_bit(MLX5E_SQ_STATE_REDIRECT, &sq->state); set_bit(MLX5E_SQ_STATE_ENABLED, &sq->state); err = mlx5e_create_sq_rdy(c->mdev, param, &csp, &sq->sqn); if (err) @@ -1612,7 +1610,7 @@ err_free_xdpsq: return err; } -static void mlx5e_close_xdpsq(struct mlx5e_xdpsq *sq) +static void mlx5e_close_xdpsq(struct mlx5e_xdpsq *sq, struct mlx5e_rq *rq) { struct mlx5e_channel *c = sq->channel; @@ -1620,7 +1618,7 @@ static void mlx5e_close_xdpsq(struct mlx5e_xdpsq *sq) napi_synchronize(&c->napi); mlx5e_destroy_sq(c->mdev, sq->sqn); - mlx5e_free_xdpsq_descs(sq); + mlx5e_free_xdpsq_descs(sq, rq); mlx5e_free_xdpsq(sq); } @@ -2008,7 +2006,7 @@ err_close_rq: err_close_xdp_sq: if (c->xdp) - mlx5e_close_xdpsq(&c->rq.xdpsq); + mlx5e_close_xdpsq(&c->rq.xdpsq, &c->rq); err_close_sqs: mlx5e_close_sqs(c); @@ -2061,10 +2059,10 @@ static void mlx5e_deactivate_channel(struct mlx5e_channel *c) static void mlx5e_close_channel(struct mlx5e_channel *c) { - mlx5e_close_xdpsq(&c->xdpsq); + mlx5e_close_xdpsq(&c->xdpsq, NULL); mlx5e_close_rq(&c->rq); if (c->xdp) - mlx5e_close_xdpsq(&c->rq.xdpsq); + mlx5e_close_xdpsq(&c->rq.xdpsq, &c->rq); mlx5e_close_sqs(c); mlx5e_close_icosq(&c->icosq); napi_disable(&c->napi); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c index 85d517360157..b4af5e19f6ac 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c @@ -76,6 +76,7 @@ int mlx5e_napi_poll(struct napi_struct *napi, int budget) struct mlx5e_channel *c = container_of(napi, struct mlx5e_channel, napi); struct mlx5e_ch_stats *ch_stats = c->stats; + struct mlx5e_rq *rq = &c->rq; bool busy = false; int work_done = 0; int i; @@ -85,17 +86,17 @@ int mlx5e_napi_poll(struct napi_struct *napi, int budget) for (i = 0; i < c->num_tc; i++) busy |= mlx5e_poll_tx_cq(&c->sq[i].cq, budget); - busy |= mlx5e_poll_xdpsq_cq(&c->xdpsq.cq); + busy |= mlx5e_poll_xdpsq_cq(&c->xdpsq.cq, NULL); if (c->xdp) - busy |= mlx5e_poll_xdpsq_cq(&c->rq.xdpsq.cq); + busy |= mlx5e_poll_xdpsq_cq(&rq->xdpsq.cq, rq); if (likely(budget)) { /* budget=0 means: don't poll rx rings */ - work_done = mlx5e_poll_rx_cq(&c->rq.cq, budget); + work_done = mlx5e_poll_rx_cq(&rq->cq, budget); busy |= work_done == budget; } - busy |= c->rq.post_wqes(&c->rq); + busy |= c->rq.post_wqes(rq); if (busy) { if (likely(mlx5e_channel_no_affinity_change(c))) @@ -115,9 +116,9 @@ int mlx5e_napi_poll(struct napi_struct *napi, int budget) mlx5e_cq_arm(&c->sq[i].cq); } - mlx5e_handle_rx_dim(&c->rq); + mlx5e_handle_rx_dim(rq); - mlx5e_cq_arm(&c->rq.cq); + mlx5e_cq_arm(&rq->cq); mlx5e_cq_arm(&c->icosq.cq); mlx5e_cq_arm(&c->xdpsq.cq); -- cgit v1.2.3-55-g7522 From db02a308cd724fd37416012c8077d3e0ea2bf386 Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Thu, 15 Nov 2018 11:48:41 +0200 Subject: net/mlx5e: XDP, Warn upon polling an error CQE Do not ignore the CQE opcode. This helps expose issues and debug them. Signed-off-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c index 4961b8fea111..407e285b0430 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c @@ -197,6 +197,11 @@ bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq, struct mlx5e_rq *rq) wqe_counter = be16_to_cpu(cqe->wqe_counter); + if (unlikely(get_cqe_opcode(cqe) != MLX5_CQE_REQ)) + netdev_WARN_ONCE(sq->channel->netdev, + "Bad OP in XDPSQ CQE: 0x%x\n", + get_cqe_opcode(cqe)); + do { u16 ci = mlx5_wq_cyc_ctr2ix(&sq->wq, sqcc); struct mlx5e_xdp_info *xdpi = &sq->db.xdpi[ci]; -- cgit v1.2.3-55-g7522 From b8180392edd97cd5bfdf12270315a72fe601cf7e Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Wed, 21 Nov 2018 14:06:02 +0200 Subject: net/mlx5e: XDP, Replace boolean doorbell indication with segment pointer Instead of calculating the control segment to be used upon an XDP xmit doorbell, save it in SQ structure. Nullify when no pending doorbell. Signed-off-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 2 +- drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c | 14 ++++---------- drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h | 11 ++++------- 3 files changed, 9 insertions(+), 18 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 15cbf0c6e309..35cd4099feda 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -413,7 +413,7 @@ struct mlx5e_xdpsq { /* dirtied @xmit */ u16 pc ____cacheline_aligned_in_smp; - bool doorbell; + struct mlx5_wqe_ctrl_seg *doorbell_cseg; 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 407e285b0430..fb1cdf729efb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c @@ -126,11 +126,8 @@ bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi) } if (unlikely(!mlx5e_wqc_has_room_for(wq, sq->cc, sq->pc, 1))) { - if (sq->doorbell) { - /* SQ is full, ring doorbell */ - mlx5e_xmit_xdp_doorbell(sq); - sq->doorbell = false; - } + /* SQ is full, ring doorbell */ + mlx5e_xmit_xdp_doorbell(sq); stats->full++; return false; } @@ -158,7 +155,7 @@ bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi) sq->db.xdpi[pi] = *xdpi; sq->pc++; - sq->doorbell = true; + sq->doorbell_cseg = cseg; stats->xmit++; return true; @@ -309,10 +306,7 @@ void mlx5e_xdp_rx_poll_complete(struct mlx5e_rq *rq) { struct mlx5e_xdpsq *xdpsq = &rq->xdpsq; - if (xdpsq->doorbell) { - mlx5e_xmit_xdp_doorbell(xdpsq); - xdpsq->doorbell = false; - } + mlx5e_xmit_xdp_doorbell(xdpsq); if (xdpsq->redirect_flush) { xdp_do_flush_map(); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h index 37fcb17e7f27..e2faf869e77c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h @@ -51,13 +51,10 @@ int mlx5e_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames, static inline void mlx5e_xmit_xdp_doorbell(struct mlx5e_xdpsq *sq) { - struct mlx5_wq_cyc *wq = &sq->wq; - struct mlx5e_tx_wqe *wqe; - u16 pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc - 1); /* last pi */ - - wqe = mlx5_wq_cyc_get_wqe(wq, pi); - - mlx5e_notify_hw(wq, sq->pc, sq->uar_map, &wqe->ctrl); + if (sq->doorbell_cseg) { + mlx5e_notify_hw(&sq->wq, sq->pc, sq->uar_map, sq->doorbell_cseg); + sq->doorbell_cseg = NULL; + } } #endif -- cgit v1.2.3-55-g7522 From fea28dd6a281045e18c1412ab5bba54436c11088 Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Sun, 14 Oct 2018 14:37:48 +0300 Subject: net/mlx5e: XDP, Maintain a FIFO structure for xdp_info instances This provides infrastructure to have multiple xdp_info instances for the same consumer index. Signed-off-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 13 +++++++-- drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c | 32 +++++++++++------------ drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h | 15 +++++++++++ drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 29 +++++++++++++++----- 4 files changed, 65 insertions(+), 24 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 35cd4099feda..fab49cbdb85d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -404,15 +404,24 @@ struct mlx5e_xdp_info { struct mlx5e_dma_info di; }; +struct mlx5e_xdp_info_fifo { + struct mlx5e_xdp_info *xi; + u32 *cc; + u32 *pc; + u32 mask; +}; + struct mlx5e_xdpsq { /* data path */ /* dirtied @completion */ + u32 xdpi_fifo_cc; u16 cc; bool redirect_flush; /* dirtied @xmit */ - u16 pc ____cacheline_aligned_in_smp; + u32 xdpi_fifo_pc ____cacheline_aligned_in_smp; + u16 pc; struct mlx5_wqe_ctrl_seg *doorbell_cseg; struct mlx5e_cq cq; @@ -421,7 +430,7 @@ struct mlx5e_xdpsq { struct mlx5_wq_cyc wq; struct mlx5e_xdpsq_stats *stats; struct { - struct mlx5e_xdp_info *xdpi; + struct mlx5e_xdp_info_fifo xdpi_fifo; } db; void __iomem *uar_map; u32 sqn; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c index fb1cdf729efb..fe8788a7744e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c @@ -149,20 +149,18 @@ bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi) cseg->opmod_idx_opcode = cpu_to_be32((sq->pc << 8) | MLX5_OPCODE_SEND); - /* move page to reference to sq responsibility, - * and mark so it's not put back in page-cache. - */ - sq->db.xdpi[pi] = *xdpi; sq->pc++; sq->doorbell_cseg = cseg; + mlx5e_xdpi_fifo_push(&sq->db.xdpi_fifo, xdpi); stats->xmit++; return true; } bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq, struct mlx5e_rq *rq) { + struct mlx5e_xdp_info_fifo *xdpi_fifo; struct mlx5e_xdpsq *sq; struct mlx5_cqe64 *cqe; bool is_redirect; @@ -179,6 +177,7 @@ bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq, struct mlx5e_rq *rq) return false; is_redirect = !rq; + xdpi_fifo = &sq->db.xdpi_fifo; /* sq->cc must be updated only after mlx5_cqwq_update_db_record(), * otherwise a cq overrun may occur @@ -200,19 +199,19 @@ bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq, struct mlx5e_rq *rq) get_cqe_opcode(cqe)); do { - u16 ci = mlx5_wq_cyc_ctr2ix(&sq->wq, sqcc); - struct mlx5e_xdp_info *xdpi = &sq->db.xdpi[ci]; + struct mlx5e_xdp_info xdpi = + mlx5e_xdpi_fifo_pop(xdpi_fifo); last_wqe = (sqcc == wqe_counter); sqcc++; if (is_redirect) { - xdp_return_frame(xdpi->xdpf); - dma_unmap_single(sq->pdev, xdpi->dma_addr, - xdpi->xdpf->len, DMA_TO_DEVICE); + xdp_return_frame(xdpi.xdpf); + dma_unmap_single(sq->pdev, xdpi.dma_addr, + xdpi.xdpf->len, DMA_TO_DEVICE); } else { /* Recycle RX page */ - mlx5e_page_release(rq, &xdpi->di, true); + mlx5e_page_release(rq, &xdpi.di, true); } } while (!last_wqe); } while ((++i < MLX5E_TX_CQ_POLL_BUDGET) && (cqe = mlx5_cqwq_get_cqe(&cq->wq))); @@ -230,21 +229,22 @@ bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq, struct mlx5e_rq *rq) void mlx5e_free_xdpsq_descs(struct mlx5e_xdpsq *sq, struct mlx5e_rq *rq) { + struct mlx5e_xdp_info_fifo *xdpi_fifo = &sq->db.xdpi_fifo; bool is_redirect = !rq; while (sq->cc != sq->pc) { - u16 ci = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->cc); - struct mlx5e_xdp_info *xdpi = &sq->db.xdpi[ci]; + struct mlx5e_xdp_info xdpi = + mlx5e_xdpi_fifo_pop(xdpi_fifo); sq->cc++; if (is_redirect) { - xdp_return_frame(xdpi->xdpf); - dma_unmap_single(sq->pdev, xdpi->dma_addr, - xdpi->xdpf->len, DMA_TO_DEVICE); + xdp_return_frame(xdpi.xdpf); + dma_unmap_single(sq->pdev, xdpi.dma_addr, + xdpi.xdpf->len, DMA_TO_DEVICE); } else { /* Recycle RX page */ - mlx5e_page_release(rq, &xdpi->di, false); + mlx5e_page_release(rq, &xdpi.di, false); } } } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h index e2faf869e77c..fd689ed506af 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h @@ -57,4 +57,19 @@ static inline void mlx5e_xmit_xdp_doorbell(struct mlx5e_xdpsq *sq) } } +static inline void +mlx5e_xdpi_fifo_push(struct mlx5e_xdp_info_fifo *fifo, + struct mlx5e_xdp_info *xi) +{ + u32 i = (*fifo->pc)++ & fifo->mask; + + fifo->xi[i] = *xi; +} + +static inline struct mlx5e_xdp_info +mlx5e_xdpi_fifo_pop(struct mlx5e_xdp_info_fifo *fifo) +{ + return fifo->xi[(*fifo->cc)++ & fifo->mask]; +} + #endif diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 8a19f451fb7a..a9f773f6d09a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -992,18 +992,35 @@ static void mlx5e_close_rq(struct mlx5e_rq *rq) static void mlx5e_free_xdpsq_db(struct mlx5e_xdpsq *sq) { - kvfree(sq->db.xdpi); + kvfree(sq->db.xdpi_fifo.xi); +} + +static int mlx5e_alloc_xdpsq_fifo(struct mlx5e_xdpsq *sq, int numa) +{ + struct mlx5e_xdp_info_fifo *xdpi_fifo = &sq->db.xdpi_fifo; + int wq_sz = mlx5_wq_cyc_get_size(&sq->wq); + int dsegs_per_wq = wq_sz * MLX5_SEND_WQEBB_NUM_DS; + + xdpi_fifo->xi = kvzalloc_node(sizeof(*xdpi_fifo->xi) * dsegs_per_wq, + GFP_KERNEL, numa); + if (!xdpi_fifo->xi) + return -ENOMEM; + + xdpi_fifo->pc = &sq->xdpi_fifo_pc; + xdpi_fifo->cc = &sq->xdpi_fifo_cc; + xdpi_fifo->mask = dsegs_per_wq - 1; + + return 0; } static int mlx5e_alloc_xdpsq_db(struct mlx5e_xdpsq *sq, int numa) { - int wq_sz = mlx5_wq_cyc_get_size(&sq->wq); + int err; - sq->db.xdpi = kvzalloc_node(array_size(wq_sz, sizeof(*sq->db.xdpi)), - GFP_KERNEL, numa); - if (!sq->db.xdpi) { + err = mlx5e_alloc_xdpsq_fifo(sq, numa); + if (err) { mlx5e_free_xdpsq_db(sq); - return -ENOMEM; + return err; } return 0; -- cgit v1.2.3-55-g7522 From 1feeab80078078186b5080fe15c51a5cce724a0e Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Sun, 14 Oct 2018 14:46:57 +0300 Subject: net/mlx5e: XDP, Add array for WQE info descriptors Each xdp_wqe_info instance describes the number of data-segments and WQEBBs of the WQE. This is useful for a downstream patch that adds support for Multi-Packet TX WQE feature. Signed-off-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 6 +++ drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c | 58 +++++++++++++++-------- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 11 +++++ 3 files changed, 54 insertions(+), 21 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index fab49cbdb85d..20b3432b35de 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -411,6 +411,11 @@ struct mlx5e_xdp_info_fifo { u32 mask; }; +struct mlx5e_xdp_wqe_info { + u8 num_wqebbs; + u8 num_ds; +}; + struct mlx5e_xdpsq { /* data path */ @@ -430,6 +435,7 @@ struct mlx5e_xdpsq { struct mlx5_wq_cyc wq; struct mlx5e_xdpsq_stats *stats; struct { + struct mlx5e_xdp_wqe_info *wqe_info; struct mlx5e_xdp_info_fifo xdpi_fifo; } db; void __iomem *uar_map; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c index fe8788a7744e..5e5e43ea9b53 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c @@ -199,19 +199,27 @@ bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq, struct mlx5e_rq *rq) get_cqe_opcode(cqe)); do { - struct mlx5e_xdp_info xdpi = - mlx5e_xdpi_fifo_pop(xdpi_fifo); + struct mlx5e_xdp_wqe_info *wi; + u16 ci, j; last_wqe = (sqcc == wqe_counter); - sqcc++; - - if (is_redirect) { - xdp_return_frame(xdpi.xdpf); - dma_unmap_single(sq->pdev, xdpi.dma_addr, - xdpi.xdpf->len, DMA_TO_DEVICE); - } else { - /* Recycle RX page */ - mlx5e_page_release(rq, &xdpi.di, true); + ci = mlx5_wq_cyc_ctr2ix(&sq->wq, sqcc); + wi = &sq->db.wqe_info[ci]; + + sqcc += wi->num_wqebbs; + + for (j = 0; j < wi->num_ds; j++) { + struct mlx5e_xdp_info xdpi = + mlx5e_xdpi_fifo_pop(xdpi_fifo); + + if (is_redirect) { + xdp_return_frame(xdpi.xdpf); + dma_unmap_single(sq->pdev, xdpi.dma_addr, + xdpi.xdpf->len, DMA_TO_DEVICE); + } else { + /* Recycle RX page */ + mlx5e_page_release(rq, &xdpi.di, true); + } } } while (!last_wqe); } while ((++i < MLX5E_TX_CQ_POLL_BUDGET) && (cqe = mlx5_cqwq_get_cqe(&cq->wq))); @@ -233,18 +241,26 @@ void mlx5e_free_xdpsq_descs(struct mlx5e_xdpsq *sq, struct mlx5e_rq *rq) bool is_redirect = !rq; while (sq->cc != sq->pc) { - struct mlx5e_xdp_info xdpi = - mlx5e_xdpi_fifo_pop(xdpi_fifo); + struct mlx5e_xdp_wqe_info *wi; + u16 ci, i; - sq->cc++; + ci = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->cc); + wi = &sq->db.wqe_info[ci]; - if (is_redirect) { - xdp_return_frame(xdpi.xdpf); - dma_unmap_single(sq->pdev, xdpi.dma_addr, - xdpi.xdpf->len, DMA_TO_DEVICE); - } else { - /* Recycle RX page */ - mlx5e_page_release(rq, &xdpi.di, false); + sq->cc += wi->num_wqebbs; + + for (i = 0; i < wi->num_ds; i++) { + struct mlx5e_xdp_info xdpi = + mlx5e_xdpi_fifo_pop(xdpi_fifo); + + if (is_redirect) { + xdp_return_frame(xdpi.xdpf); + dma_unmap_single(sq->pdev, xdpi.dma_addr, + xdpi.xdpf->len, DMA_TO_DEVICE); + } else { + /* Recycle RX page */ + mlx5e_page_release(rq, &xdpi.di, false); + } } } } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index a9f773f6d09a..7086c73d5915 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -993,6 +993,7 @@ static void mlx5e_close_rq(struct mlx5e_rq *rq) static void mlx5e_free_xdpsq_db(struct mlx5e_xdpsq *sq) { kvfree(sq->db.xdpi_fifo.xi); + kvfree(sq->db.wqe_info); } static int mlx5e_alloc_xdpsq_fifo(struct mlx5e_xdpsq *sq, int numa) @@ -1015,8 +1016,14 @@ static int mlx5e_alloc_xdpsq_fifo(struct mlx5e_xdpsq *sq, int numa) static int mlx5e_alloc_xdpsq_db(struct mlx5e_xdpsq *sq, int numa) { + int wq_sz = mlx5_wq_cyc_get_size(&sq->wq); int err; + sq->db.wqe_info = kvzalloc_node(sizeof(*sq->db.wqe_info) * wq_sz, + GFP_KERNEL, numa); + if (!sq->db.wqe_info) + return -ENOMEM; + err = mlx5e_alloc_xdpsq_fifo(sq, numa); if (err) { mlx5e_free_xdpsq_db(sq); @@ -1606,6 +1613,7 @@ static int mlx5e_open_xdpsq(struct mlx5e_channel *c, /* Pre initialize fixed WQE fields */ for (i = 0; i < mlx5_wq_cyc_get_size(&sq->wq); i++) { + struct mlx5e_xdp_wqe_info *wi = &sq->db.wqe_info[i]; struct mlx5e_tx_wqe *wqe = mlx5_wq_cyc_get_wqe(&sq->wq, i); struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl; struct mlx5_wqe_eth_seg *eseg = &wqe->eth; @@ -1616,6 +1624,9 @@ static int mlx5e_open_xdpsq(struct mlx5e_channel *c, dseg = (struct mlx5_wqe_data_seg *)cseg + (ds_cnt - 1); dseg->lkey = sq->mkey_be; + + wi->num_wqebbs = 1; + wi->num_ds = 1; } return 0; -- cgit v1.2.3-55-g7522 From 5e0d2eef771ee78b092bf93d040eac02a0965fea Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Wed, 21 Nov 2018 14:08:06 +0200 Subject: net/mlx5e: XDP, Support Enhanced Multi-Packet TX WQE Add support for the HW feature of multi-packet WQE in XDP xmit flow. The conventional TX descriptor (WQE, Work Queue Element) serves a single packet. Our HW has support for multi-packet WQE (MPWQE) in which a single descriptor serves multiple TX packets. This reduces both the PCI overhead and the CPU cycles wasted on writing them. In this patch we add support for the HW feature, which is supported starting from ConnectX-5. Performance: Tested packet rate for UDP 64Byte multi-stream over ConnectX-5 NICs. CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz XDP_TX: We see a huge gain on single port ConnectX-5, and reach the 100 Mpps milestone. * Single-port HCA: Before: 70 Mpps After: 100 Mpps (+42.8%) * Dual-port HCA: Before: 51.7 Mpps After: 57.3 Mpps (+10.8%) * In both cases we tested traffic on one port and for now On Dual-port HCAs we see only small gain, we are working to overcome this bottleneck, but for the moment only with experimental firmware on dual port HCAs we can reach the wanted numbers as seen on Single-port HCAs. XDP_REDIRECT: Redirect from (A) ConnectX-5 to (B) ConnectX-5. Due to a setup limitation, (A) and (B) are on different NUMA nodes, so absolute performance numbers are not optimal. Note: Below is the transmit rate of (B), not the redirect rate of (A) which is in some cases higher. * (B) is single-port: Before: 77 Mpps After: 90 Mpps (+16.8%) * (B) is dual-port: Before: 61 Mpps After: 72 Mpps (+18%) Signed-off-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 12 +++ drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c | 112 +++++++++++++++++++++- drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h | 29 +++++- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 47 +++++---- include/linux/mlx5/device.h | 1 + 5 files changed, 174 insertions(+), 27 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 20b3432b35de..8f5545d317ba 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -416,6 +416,16 @@ struct mlx5e_xdp_wqe_info { u8 num_ds; }; +struct mlx5e_xdp_mpwqe { + /* Current MPWQE session */ + struct mlx5e_tx_wqe *wqe; + u8 ds_count; + u8 max_ds_count; +}; + +struct mlx5e_xdpsq; +typedef bool (*mlx5e_fp_xmit_xdp_frame)(struct mlx5e_xdpsq*, + struct mlx5e_xdp_info*); struct mlx5e_xdpsq { /* data path */ @@ -428,12 +438,14 @@ struct mlx5e_xdpsq { u32 xdpi_fifo_pc ____cacheline_aligned_in_smp; u16 pc; struct mlx5_wqe_ctrl_seg *doorbell_cseg; + struct mlx5e_xdp_mpwqe mpwqe; struct mlx5e_cq cq; /* read only */ struct mlx5_wq_cyc wq; struct mlx5e_xdpsq_stats *stats; + mlx5e_fp_xmit_xdp_frame xmit_xdp_frame; struct { struct mlx5e_xdp_wqe_info *wqe_info; struct mlx5e_xdp_info_fifo xdpi_fifo; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c index 5e5e43ea9b53..3740177eed09 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c @@ -47,7 +47,7 @@ mlx5e_xmit_xdp_buff(struct mlx5e_xdpsq *sq, struct mlx5e_dma_info *di, xdpi.xdpf->len, PCI_DMA_TODEVICE); xdpi.di = *di; - return mlx5e_xmit_xdp_frame(sq, &xdpi); + return sq->xmit_xdp_frame(sq, &xdpi); } /* returns true if packet was consumed by xdp */ @@ -102,7 +102,98 @@ xdp_abort: } } -bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi) +static void mlx5e_xdp_mpwqe_session_start(struct mlx5e_xdpsq *sq) +{ + struct mlx5e_xdp_mpwqe *session = &sq->mpwqe; + 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; + + pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc); + +/* 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 + + wqebbs = min_t(u16, mlx5_wq_cyc_get_contig_wqebbs(wq, pi), + MLX5E_XDP_MPW_MAX_WQEBBS); + + session->max_ds_count = MLX5_SEND_WQEBB_NUM_DS * wqebbs; +} + +static void mlx5e_xdp_mpwqe_complete(struct mlx5e_xdpsq *sq) +{ + struct mlx5_wq_cyc *wq = &sq->wq; + struct mlx5e_xdp_mpwqe *session = &sq->mpwqe; + struct mlx5_wqe_ctrl_seg *cseg = &session->wqe->ctrl; + u16 ds_count = session->ds_count; + u16 pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc); + struct mlx5e_xdp_wqe_info *wi = &sq->db.wqe_info[pi]; + + cseg->opmod_idx_opcode = + cpu_to_be32((sq->pc << 8) | MLX5_OPCODE_ENHANCED_MPSW); + cseg->qpn_ds = cpu_to_be32((sq->sqn << 8) | ds_count); + + wi->num_wqebbs = DIV_ROUND_UP(ds_count, MLX5_SEND_WQEBB_NUM_DS); + wi->num_ds = ds_count - MLX5E_XDP_TX_EMPTY_DS_COUNT; + + sq->pc += wi->num_wqebbs; + + sq->doorbell_cseg = cseg; + + session->wqe = NULL; /* Close session */ +} + +static bool mlx5e_xmit_xdp_frame_mpwqe(struct mlx5e_xdpsq *sq, + struct mlx5e_xdp_info *xdpi) +{ + struct mlx5e_xdp_mpwqe *session = &sq->mpwqe; + struct mlx5e_xdpsq_stats *stats = sq->stats; + + dma_addr_t dma_addr = xdpi->dma_addr; + struct xdp_frame *xdpf = xdpi->xdpf; + unsigned int dma_len = xdpf->len; + + if (unlikely(sq->hw_mtu < dma_len)) { + stats->err++; + return false; + } + + if (unlikely(!session->wqe)) { + if (unlikely(!mlx5e_wqc_has_room_for(&sq->wq, sq->cc, sq->pc, + MLX5_SEND_WQE_MAX_WQEBBS))) { + /* SQ is full, ring doorbell */ + mlx5e_xmit_xdp_doorbell(sq); + stats->full++; + return false; + } + + mlx5e_xdp_mpwqe_session_start(sq); + } + + mlx5e_xdp_mpwqe_add_dseg(sq, dma_addr, dma_len); + + if (unlikely(session->ds_count == session->max_ds_count)) + mlx5e_xdp_mpwqe_complete(sq); + + mlx5e_xdpi_fifo_push(&sq->db.xdpi_fifo, xdpi); + stats->xmit++; + return true; +} + +static bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi) { struct mlx5_wq_cyc *wq = &sq->wq; u16 pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc); @@ -304,7 +395,7 @@ int mlx5e_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames, xdpi.xdpf = xdpf; - if (unlikely(!mlx5e_xmit_xdp_frame(sq, &xdpi))) { + if (unlikely(!sq->xmit_xdp_frame(sq, &xdpi))) { dma_unmap_single(sq->pdev, xdpi.dma_addr, xdpf->len, DMA_TO_DEVICE); xdp_return_frame_rx_napi(xdpf); @@ -312,8 +403,11 @@ int mlx5e_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames, } } - if (flags & XDP_XMIT_FLUSH) + if (flags & XDP_XMIT_FLUSH) { + if (sq->mpwqe.wqe) + mlx5e_xdp_mpwqe_complete(sq); mlx5e_xmit_xdp_doorbell(sq); + } return n - drops; } @@ -322,6 +416,9 @@ void mlx5e_xdp_rx_poll_complete(struct mlx5e_rq *rq) { struct mlx5e_xdpsq *xdpsq = &rq->xdpsq; + if (xdpsq->mpwqe.wqe) + mlx5e_xdp_mpwqe_complete(xdpsq); + mlx5e_xmit_xdp_doorbell(xdpsq); if (xdpsq->redirect_flush) { @@ -329,3 +426,10 @@ void mlx5e_xdp_rx_poll_complete(struct mlx5e_rq *rq) xdpsq->redirect_flush = false; } } + +void mlx5e_set_xmit_fp(struct mlx5e_xdpsq *sq, bool is_mpw) +{ + sq->xmit_xdp_frame = is_mpw ? + mlx5e_xmit_xdp_frame_mpwqe : mlx5e_xmit_xdp_frame; +} + diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h index fd689ed506af..3a67cb3cd179 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h @@ -37,15 +37,16 @@ #define MLX5E_XDP_MAX_MTU ((int)(PAGE_SIZE - \ MLX5_SKB_FRAG_SZ(XDP_PACKET_HEADROOM))) #define MLX5E_XDP_MIN_INLINE (ETH_HLEN + VLAN_HLEN) -#define MLX5E_XDP_TX_DS_COUNT \ - ((sizeof(struct mlx5e_tx_wqe) / MLX5_SEND_WQE_DS) + 1 /* SG DS */) +#define MLX5E_XDP_TX_EMPTY_DS_COUNT \ + (sizeof(struct mlx5e_tx_wqe) / MLX5_SEND_WQE_DS) +#define MLX5E_XDP_TX_DS_COUNT (MLX5E_XDP_TX_EMPTY_DS_COUNT + 1 /* SG DS */) bool mlx5e_xdp_handle(struct mlx5e_rq *rq, struct mlx5e_dma_info *di, void *va, u16 *rx_headroom, u32 *len); bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq, struct mlx5e_rq *rq); void mlx5e_free_xdpsq_descs(struct mlx5e_xdpsq *sq, struct mlx5e_rq *rq); +void mlx5e_set_xmit_fp(struct mlx5e_xdpsq *sq, bool is_mpw); void mlx5e_xdp_rx_poll_complete(struct mlx5e_rq *rq); -bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi); int mlx5e_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames, u32 flags); @@ -57,6 +58,28 @@ static inline void mlx5e_xmit_xdp_doorbell(struct mlx5e_xdpsq *sq) } } +static inline void +mlx5e_xdp_mpwqe_add_dseg(struct mlx5e_xdpsq *sq, dma_addr_t dma_addr, u16 dma_len) +{ + struct mlx5e_xdp_mpwqe *session = &sq->mpwqe; + struct mlx5_wqe_data_seg *dseg = + (struct mlx5_wqe_data_seg *)session->wqe + session->ds_count++; + + dseg->addr = cpu_to_be64(dma_addr); + dseg->byte_count = cpu_to_be32(dma_len); + dseg->lkey = sq->mkey_be; +} + +static inline void mlx5e_xdpsq_fetch_wqe(struct mlx5e_xdpsq *sq, + struct mlx5e_tx_wqe **wqe) +{ + struct mlx5_wq_cyc *wq = &sq->wq; + u16 pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc); + + *wqe = mlx5_wq_cyc_get_wqe(wq, pi); + memset(*wqe, 0, sizeof(**wqe)); +} + static inline void mlx5e_xdpi_fifo_push(struct mlx5e_xdp_info_fifo *fifo, struct mlx5e_xdp_info *xi) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 7086c73d5915..07b16e5f02bd 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -61,6 +61,7 @@ struct mlx5e_rq_param { struct mlx5e_sq_param { u32 sqc[MLX5_ST_SZ_DW(sqc)]; struct mlx5_wq_param wq; + bool is_mpw; }; struct mlx5e_cq_param { @@ -1586,11 +1587,8 @@ static int mlx5e_open_xdpsq(struct mlx5e_channel *c, struct mlx5e_xdpsq *sq, bool is_redirect) { - unsigned int ds_cnt = MLX5E_XDP_TX_DS_COUNT; struct mlx5e_create_sq_param csp = {}; - unsigned int inline_hdr_sz = 0; int err; - int i; err = mlx5e_alloc_xdpsq(c, params, param, sq, is_redirect); if (err) @@ -1606,27 +1604,35 @@ static int mlx5e_open_xdpsq(struct mlx5e_channel *c, if (err) goto err_free_xdpsq; - if (sq->min_inline_mode != MLX5_INLINE_MODE_NONE) { - inline_hdr_sz = MLX5E_XDP_MIN_INLINE; - ds_cnt++; - } + mlx5e_set_xmit_fp(sq, param->is_mpw); + + if (!param->is_mpw) { + unsigned int ds_cnt = MLX5E_XDP_TX_DS_COUNT; + unsigned int inline_hdr_sz = 0; + int i; - /* Pre initialize fixed WQE fields */ - for (i = 0; i < mlx5_wq_cyc_get_size(&sq->wq); i++) { - struct mlx5e_xdp_wqe_info *wi = &sq->db.wqe_info[i]; - struct mlx5e_tx_wqe *wqe = mlx5_wq_cyc_get_wqe(&sq->wq, i); - struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl; - struct mlx5_wqe_eth_seg *eseg = &wqe->eth; - struct mlx5_wqe_data_seg *dseg; + if (sq->min_inline_mode != MLX5_INLINE_MODE_NONE) { + inline_hdr_sz = MLX5E_XDP_MIN_INLINE; + ds_cnt++; + } + + /* Pre initialize fixed WQE fields */ + for (i = 0; i < mlx5_wq_cyc_get_size(&sq->wq); i++) { + struct mlx5e_xdp_wqe_info *wi = &sq->db.wqe_info[i]; + struct mlx5e_tx_wqe *wqe = mlx5_wq_cyc_get_wqe(&sq->wq, i); + struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl; + struct mlx5_wqe_eth_seg *eseg = &wqe->eth; + struct mlx5_wqe_data_seg *dseg; - cseg->qpn_ds = cpu_to_be32((sq->sqn << 8) | ds_cnt); - eseg->inline_hdr.sz = cpu_to_be16(inline_hdr_sz); + cseg->qpn_ds = cpu_to_be32((sq->sqn << 8) | ds_cnt); + eseg->inline_hdr.sz = cpu_to_be16(inline_hdr_sz); - dseg = (struct mlx5_wqe_data_seg *)cseg + (ds_cnt - 1); - dseg->lkey = sq->mkey_be; + dseg = (struct mlx5_wqe_data_seg *)cseg + (ds_cnt - 1); + dseg->lkey = sq->mkey_be; - wi->num_wqebbs = 1; - wi->num_ds = 1; + wi->num_wqebbs = 1; + wi->num_ds = 1; + } } return 0; @@ -2335,6 +2341,7 @@ static void mlx5e_build_xdpsq_param(struct mlx5e_priv *priv, mlx5e_build_sq_param_common(priv, param); MLX5_SET(wq, wq, log_wq_sz, params->log_sq_size); + param->is_mpw = MLX5_CAP_ETH(priv->mdev, enhanced_multi_pkt_send_wqe); } static void mlx5e_build_channel_param(struct mlx5e_priv *priv, diff --git a/include/linux/mlx5/device.h b/include/linux/mlx5/device.h index 4674b9e99f45..8c4a820bd4c1 100644 --- a/include/linux/mlx5/device.h +++ b/include/linux/mlx5/device.h @@ -421,6 +421,7 @@ enum { MLX5_OPCODE_ATOMIC_MASKED_FA = 0x15, MLX5_OPCODE_BIND_MW = 0x18, MLX5_OPCODE_CONFIG_CMD = 0x1f, + MLX5_OPCODE_ENHANCED_MPSW = 0x29, MLX5_RECV_OPCODE_RDMA_WRITE_IMM = 0x00, MLX5_RECV_OPCODE_SEND = 0x01, -- cgit v1.2.3-55-g7522 From 6277053afa34253eb8d8d147388e4426e615bae4 Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Tue, 20 Nov 2018 11:50:30 +0200 Subject: net/mlx5e: XDP, Add user control for XDP TX MPWQE feature Add ethtool private flag 'xdp_tx_mpwqe' to control the feature from userspace. Feature is set ON by default, if supported. Signed-off-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en.h | 1 + .../net/ethernet/mellanox/mlx5/core/en_ethtool.c | 28 ++++++++++++++++++++++ drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 6 ++++- 3 files changed, 34 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 8f5545d317ba..8fa8fdd30b85 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -216,6 +216,7 @@ enum mlx5e_priv_flag { MLX5E_PFLAG_RX_CQE_COMPRESS, MLX5E_PFLAG_RX_STRIDING_RQ, MLX5E_PFLAG_RX_NO_CSUM_COMPLETE, + MLX5E_PFLAG_XDP_TX_MPWQE, MLX5E_NUM_PFLAGS, /* Keep last */ }; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c index 6e101201dcbb..c9df08133718 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c @@ -1672,12 +1672,40 @@ static int set_pflag_rx_no_csum_complete(struct net_device *netdev, bool enable) return 0; } +static int set_pflag_xdp_tx_mpwqe(struct net_device *netdev, bool enable) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + struct mlx5_core_dev *mdev = priv->mdev; + struct mlx5e_channels new_channels = {}; + int err; + + if (enable && !MLX5_CAP_ETH(mdev, enhanced_multi_pkt_send_wqe)) + return -EOPNOTSUPP; + + new_channels.params = priv->channels.params; + + MLX5E_SET_PFLAG(&new_channels.params, MLX5E_PFLAG_XDP_TX_MPWQE, enable); + + if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) { + priv->channels.params = new_channels.params; + return 0; + } + + err = mlx5e_open_channels(priv, &new_channels); + if (err) + return err; + + mlx5e_switch_priv_channels(priv, &new_channels, NULL); + return 0; +} + static const struct pflag_desc mlx5e_priv_flags[MLX5E_NUM_PFLAGS] = { { "rx_cqe_moder", set_pflag_rx_cqe_based_moder }, { "tx_cqe_moder", set_pflag_tx_cqe_based_moder }, { "rx_cqe_compress", set_pflag_rx_cqe_compress }, { "rx_striding_rq", set_pflag_rx_striding_rq }, { "rx_no_csum_complete", set_pflag_rx_no_csum_complete }, + { "xdp_tx_mpwqe", set_pflag_xdp_tx_mpwqe }, }; static int mlx5e_handle_pflag(struct net_device *netdev, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 07b16e5f02bd..8cfd2ec7c0a2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -2341,7 +2341,7 @@ static void mlx5e_build_xdpsq_param(struct mlx5e_priv *priv, mlx5e_build_sq_param_common(priv, param); MLX5_SET(wq, wq, log_wq_sz, params->log_sq_size); - param->is_mpw = MLX5_CAP_ETH(priv->mdev, enhanced_multi_pkt_send_wqe); + param->is_mpw = MLX5E_GET_PFLAG(params, MLX5E_PFLAG_XDP_TX_MPWQE); } static void mlx5e_build_channel_param(struct mlx5e_priv *priv, @@ -4595,6 +4595,10 @@ void mlx5e_build_nic_params(struct mlx5_core_dev *mdev, MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE : MLX5E_PARAMS_DEFAULT_LOG_SQ_SIZE; + /* XDP SQ */ + MLX5E_SET_PFLAG(params, MLX5E_PFLAG_XDP_TX_MPWQE, + MLX5_CAP_ETH(mdev, enhanced_multi_pkt_send_wqe)); + /* set CQE compression */ params->rx_cqe_compress_def = false; if (MLX5_CAP_GEN(mdev, cqe_compression) && -- cgit v1.2.3-55-g7522 From 5d1f7354faad4b6210fac9192d0ca4c16b00ae7e Mon Sep 17 00:00:00 2001 From: kbuild test robot Date: Sat, 22 Dec 2018 20:02:16 +0800 Subject: net/mlx5e: fix semicolon.cocci warnings drivers/net/ethernet/mellanox/mlx5/core/en_rep.c:1339:57-58: Unneeded semicolon Remove unneeded semicolon. Generated by: scripts/coccinelle/misc/semicolon.cocci Fixes: 4c8fb2986d44 ("net/mlx5e: Increase VF representors' SQ size to 128") CC: Gavi Teitz Signed-off-by: kbuild test robot Reviewed-by: Leon Romanovsky Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index dbbdbdff75e8..96cc0c6a4014 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1346,7 +1346,7 @@ static void mlx5e_build_rep_params(struct net_device *netdev) if (rep->vport == FDB_UPLINK_VPORT) params->log_sq_size = MLX5E_PARAMS_DEFAULT_LOG_SQ_SIZE; else - params->log_sq_size = MLX5E_REP_PARAMS_DEF_LOG_SQ_SIZE;; + params->log_sq_size = MLX5E_REP_PARAMS_DEF_LOG_SQ_SIZE; /* RQ */ mlx5e_build_rq_params(mdev, params); -- cgit v1.2.3-55-g7522 From 2534f14a94c1fe48928a4b00a0e241fa7132d00b Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Sun, 23 Dec 2018 09:56:57 +0100 Subject: net/mlx5e: drop useless LIST_HEAD Drop LIST_HEAD where the variable it declares is never used. These became useless in 244cd96adb5f ("net_sched: remove list_head from tc_action") The semantic patch that fixes this problem is as follows: (http://coccinelle.lip6.fr/) // @@ identifier x; @@ - LIST_HEAD(x); ... when != x // Fixes: 244cd96adb5f ("net_sched: remove list_head from tc_action") Signed-off-by: Julia Lawall Reviewed-by: Leon Romanovsky Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx5') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 4d6c9757afad..cae6c6d48984 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -2183,7 +2183,6 @@ static bool modify_header_match_supported(struct mlx5_flow_spec *spec, { const struct tc_action *a; bool modify_ip_header; - LIST_HEAD(actions); u8 htype, ip_proto; void *headers_v; u16 ethertype; @@ -2272,7 +2271,6 @@ static int parse_tc_nic_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, { struct mlx5_nic_flow_attr *attr = flow->nic_attr; const struct tc_action *a; - LIST_HEAD(actions); u32 action = 0; int err, i; @@ -2511,7 +2509,6 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, struct mlx5e_rep_priv *rpriv = priv->ppriv; struct ip_tunnel_info *info = NULL; const struct tc_action *a; - LIST_HEAD(actions); bool encap = false; u32 action = 0; int err, i; -- cgit v1.2.3-55-g7522