From 3d9c5e023a0dbf3e117bb416cfefd9405bf5af0c Mon Sep 17 00:00:00 2001 From: Huy Nguyen Date: Mon, 3 Feb 2020 16:32:18 -0600 Subject: [PATCH 1/7] net/mlx5: Fix sleep while atomic in mlx5_eswitch_get_vepa rtnl_bridge_getlink is protected by rcu lock, so mlx5_eswitch_get_vepa cannot take mutex lock. Two possible issues can happen: 1. User at the same time change vepa mode via RTM_SETLINK command. 2. User at the same time change the switchdev mode via devlink netlink interface. Case 1 cannot happen because rtnl executes one message in order. Case 2 can happen but we do not expect user to change the switchdev mode when changing vepa. Even if a user does it, so he will read a value which is no longer valid. Fixes: 8da202b24913 ("net/mlx5: E-Switch, Add support for VEPA in legacy mode.") Signed-off-by: Huy Nguyen Reviewed-by: Mark Bloch Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 14 +++----------- 1 file changed, 3 insertions(+), 11 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index 5acf60b1bbfe..564d42605892 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -2452,25 +2452,17 @@ out: int mlx5_eswitch_get_vepa(struct mlx5_eswitch *esw, u8 *setting) { - int err = 0; - if (!esw) return -EOPNOTSUPP; if (!ESW_ALLOWED(esw)) return -EPERM; - mutex_lock(&esw->state_lock); - if (esw->mode != MLX5_ESWITCH_LEGACY) { - err = -EOPNOTSUPP; - goto out; - } + if (esw->mode != MLX5_ESWITCH_LEGACY) + return -EOPNOTSUPP; *setting = esw->fdb_table.legacy.vepa_uplink_rule ? 1 : 0; - -out: - mutex_unlock(&esw->state_lock); - return err; + return 0; } int mlx5_eswitch_set_vport_trust(struct mlx5_eswitch *esw, From 5ee090ed0da649b1febae2b7c285ac77d1e55a0c Mon Sep 17 00:00:00 2001 From: Aya Levin Date: Mon, 9 Dec 2019 14:08:18 +0200 Subject: [PATCH 2/7] net/mlx5e: Reset RQ doorbell counter before moving RQ state from RST to RDY Initialize RQ doorbell counters to zero prior to moving an RQ from RST to RDY state. Per HW spec, when RQ is back to RDY state, the descriptor ID on the completion is reset. The doorbell record must comply. Fixes: 8276ea1353a4 ("net/mlx5e: Report and recover from CQE with error on RQ") Signed-off-by: Aya Levin Reported-by: Tariq Toukan Reviewed-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/en/txrx.h | 8 ++++ .../net/ethernet/mellanox/mlx5/core/en_main.c | 3 ++ drivers/net/ethernet/mellanox/mlx5/core/wq.c | 39 ++++++++++++++----- drivers/net/ethernet/mellanox/mlx5/core/wq.h | 2 + 4 files changed, 43 insertions(+), 9 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h b/drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h index 7c8796d9743f..a226277b0980 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h @@ -179,6 +179,14 @@ mlx5e_tx_dma_unmap(struct device *pdev, struct mlx5e_sq_dma *dma) } } +static inline void mlx5e_rqwq_reset(struct mlx5e_rq *rq) +{ + if (rq->wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ) + mlx5_wq_ll_reset(&rq->mpwqe.wq); + else + mlx5_wq_cyc_reset(&rq->wqe.wq); +} + /* SW parser related functions */ struct mlx5e_swp_spec { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 454d3459bd8b..966983674663 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -712,6 +712,9 @@ int mlx5e_modify_rq_state(struct mlx5e_rq *rq, int curr_state, int next_state) if (!in) return -ENOMEM; + if (curr_state == MLX5_RQC_STATE_RST && next_state == MLX5_RQC_STATE_RDY) + mlx5e_rqwq_reset(rq); + rqc = MLX5_ADDR_OF(modify_rq_in, in, ctx); MLX5_SET(modify_rq_in, in, rq_state, curr_state); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/wq.c b/drivers/net/ethernet/mellanox/mlx5/core/wq.c index 02f7e4a39578..01f075fac276 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/wq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/wq.c @@ -94,6 +94,13 @@ void mlx5_wq_cyc_wqe_dump(struct mlx5_wq_cyc *wq, u16 ix, u8 nstrides) print_hex_dump(KERN_WARNING, "", DUMP_PREFIX_OFFSET, 16, 1, wqe, len, false); } +void mlx5_wq_cyc_reset(struct mlx5_wq_cyc *wq) +{ + wq->wqe_ctr = 0; + wq->cur_sz = 0; + mlx5_wq_cyc_update_db_record(wq); +} + int mlx5_wq_qp_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, void *qpc, struct mlx5_wq_qp *wq, struct mlx5_wq_ctrl *wq_ctrl) @@ -192,6 +199,19 @@ err_db_free: return err; } +static void mlx5_wq_ll_init_list(struct mlx5_wq_ll *wq) +{ + struct mlx5_wqe_srq_next_seg *next_seg; + int i; + + for (i = 0; i < wq->fbc.sz_m1; i++) { + next_seg = mlx5_wq_ll_get_wqe(wq, i); + next_seg->next_wqe_index = cpu_to_be16(i + 1); + } + next_seg = mlx5_wq_ll_get_wqe(wq, i); + wq->tail_next = &next_seg->next_wqe_index; +} + int mlx5_wq_ll_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, void *wqc, struct mlx5_wq_ll *wq, struct mlx5_wq_ctrl *wq_ctrl) @@ -199,9 +219,7 @@ int mlx5_wq_ll_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, u8 log_wq_stride = MLX5_GET(wq, wqc, log_wq_stride); u8 log_wq_sz = MLX5_GET(wq, wqc, log_wq_sz); struct mlx5_frag_buf_ctrl *fbc = &wq->fbc; - struct mlx5_wqe_srq_next_seg *next_seg; int err; - int i; err = mlx5_db_alloc_node(mdev, &wq_ctrl->db, param->db_numa_node); if (err) { @@ -220,13 +238,7 @@ int mlx5_wq_ll_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, mlx5_init_fbc(wq_ctrl->buf.frags, log_wq_stride, log_wq_sz, fbc); - for (i = 0; i < fbc->sz_m1; i++) { - next_seg = mlx5_wq_ll_get_wqe(wq, i); - next_seg->next_wqe_index = cpu_to_be16(i + 1); - } - next_seg = mlx5_wq_ll_get_wqe(wq, i); - wq->tail_next = &next_seg->next_wqe_index; - + mlx5_wq_ll_init_list(wq); wq_ctrl->mdev = mdev; return 0; @@ -237,6 +249,15 @@ err_db_free: return err; } +void mlx5_wq_ll_reset(struct mlx5_wq_ll *wq) +{ + wq->head = 0; + wq->wqe_ctr = 0; + wq->cur_sz = 0; + mlx5_wq_ll_init_list(wq); + mlx5_wq_ll_update_db_record(wq); +} + void mlx5_wq_destroy(struct mlx5_wq_ctrl *wq_ctrl) { mlx5_frag_buf_free(wq_ctrl->mdev, &wq_ctrl->buf); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/wq.h b/drivers/net/ethernet/mellanox/mlx5/core/wq.h index d9a94bc223c0..4cadc336593f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/wq.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/wq.h @@ -80,6 +80,7 @@ int mlx5_wq_cyc_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, void *wqc, struct mlx5_wq_cyc *wq, struct mlx5_wq_ctrl *wq_ctrl); void mlx5_wq_cyc_wqe_dump(struct mlx5_wq_cyc *wq, u16 ix, u8 nstrides); +void mlx5_wq_cyc_reset(struct mlx5_wq_cyc *wq); int mlx5_wq_qp_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, void *qpc, struct mlx5_wq_qp *wq, @@ -92,6 +93,7 @@ int mlx5_cqwq_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, int mlx5_wq_ll_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, void *wqc, struct mlx5_wq_ll *wq, struct mlx5_wq_ctrl *wq_ctrl); +void mlx5_wq_ll_reset(struct mlx5_wq_ll *wq); void mlx5_wq_destroy(struct mlx5_wq_ctrl *wq_ctrl); From 1ad6c43c6a7b8627240c6cc19c69e31fedc596a7 Mon Sep 17 00:00:00 2001 From: Aya Levin Date: Wed, 12 Feb 2020 15:17:25 +0200 Subject: [PATCH 3/7] net/mlx5e: Fix crash in recovery flow without devlink reporter When health reporters are not supported, recovery function is invoked directly, not via devlink health reporters. In this direct flow, the recover function input parameter was passed incorrectly and is causing a kernel oops. This patch is fixing the input parameter. Following call trace is observed on rx error health reporting. Internal error: Oops: 96000007 [#1] PREEMPT SMP Process kworker/u16:4 (pid: 4584, stack limit = 0x00000000c9e45703) Call trace: mlx5e_rx_reporter_err_rq_cqe_recover+0x30/0x164 [mlx5_core] mlx5e_health_report+0x60/0x6c [mlx5_core] mlx5e_reporter_rq_cqe_err+0x6c/0x90 [mlx5_core] mlx5e_rq_err_cqe_work+0x20/0x2c [mlx5_core] process_one_work+0x168/0x3d0 worker_thread+0x58/0x3d0 kthread+0x108/0x134 Fixes: c50de4af1d63 ("net/mlx5e: Generalize tx reporter's functionality") Signed-off-by: Aya Levin Signed-off-by: Parav Pandit Reviewed-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en/health.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/health.c b/drivers/net/ethernet/mellanox/mlx5/core/en/health.c index 3a975641f902..20b907dc1e29 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/health.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/health.c @@ -200,7 +200,7 @@ int mlx5e_health_report(struct mlx5e_priv *priv, netdev_err(priv->netdev, err_str); if (!reporter) - return err_ctx->recover(&err_ctx->ctx); + return err_ctx->recover(err_ctx->ctx); return devlink_health_report(reporter, err_str, err_ctx); } From 52d214976d4f64504c1bbb52d47b46a5a3d5ee42 Mon Sep 17 00:00:00 2001 From: Hamdan Igbaria Date: Wed, 5 Feb 2020 14:31:12 +0200 Subject: [PATCH 4/7] net/mlx5: DR, Fix matching on vport gvmi Set vport gvmi in the tag, only when source gvmi is set in the bit mask. Fixes: 26d688e3 ("net/mlx5: DR, Add Steering entry (STE) utilities") Signed-off-by: Hamdan Igbaria Reviewed-by: Alex Vesker Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c index c6c7d1defbd7..aade62a9ee5c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c @@ -2307,7 +2307,9 @@ static int dr_ste_build_src_gvmi_qpn_tag(struct mlx5dr_match_param *value, struct mlx5dr_cmd_vport_cap *vport_cap; struct mlx5dr_domain *dmn = sb->dmn; struct mlx5dr_cmd_caps *caps; + u8 *bit_mask = sb->bit_mask; u8 *tag = hw_ste->tag; + bool source_gvmi_set; DR_STE_SET_TAG(src_gvmi_qp, tag, source_qp, misc, source_sqn); @@ -2328,7 +2330,8 @@ static int dr_ste_build_src_gvmi_qpn_tag(struct mlx5dr_match_param *value, if (!vport_cap) return -EINVAL; - if (vport_cap->vport_gvmi) + source_gvmi_set = MLX5_GET(ste_src_gvmi_qp, bit_mask, source_gvmi); + if (vport_cap->vport_gvmi && source_gvmi_set) MLX5_SET(ste_src_gvmi_qp, tag, source_gvmi, vport_cap->vport_gvmi); misc->source_eswitch_owner_vhca_id = 0; From 383de108157c881074f32914b61125e299820bd2 Mon Sep 17 00:00:00 2001 From: Dmytro Linkin Date: Wed, 12 Feb 2020 11:32:39 +0200 Subject: [PATCH 5/7] net/mlx5e: Don't clear the whole vf config when switching modes There is no need to reset all vf config (except link state) between legacy and switchdev modes changes. Also, set link state to AUTO, when legacy enabled. Fixes: 3b83b6c2e024 ("net/mlx5e: Clear VF config when switching modes") Signed-off-by: Dmytro Linkin Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 6 +++++- drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 4 ++-- 2 files changed, 7 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index 564d42605892..e49acd0c5da5 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -459,12 +459,16 @@ static void esw_destroy_legacy_table(struct mlx5_eswitch *esw) static int esw_legacy_enable(struct mlx5_eswitch *esw) { - int ret; + struct mlx5_vport *vport; + int ret, i; ret = esw_create_legacy_table(esw); if (ret) return ret; + mlx5_esw_for_each_vf_vport(esw, i, vport, esw->esw_funcs.num_vfs) + vport->info.link_state = MLX5_VPORT_ADMIN_STATE_AUTO; + ret = mlx5_eswitch_enable_pf_vf_vports(esw, MLX5_LEGACY_SRIOV_VPORT_EVENTS); if (ret) esw_destroy_legacy_table(esw); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 979f13bdc203..1a57b2bd74b8 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -1172,7 +1172,7 @@ static int esw_offloads_start(struct mlx5_eswitch *esw, return -EINVAL; } - mlx5_eswitch_disable(esw, true); + mlx5_eswitch_disable(esw, false); mlx5_eswitch_update_num_of_vfs(esw, esw->dev->priv.sriov.num_vfs); err = mlx5_eswitch_enable(esw, MLX5_ESWITCH_OFFLOADS); if (err) { @@ -2065,7 +2065,7 @@ static int esw_offloads_stop(struct mlx5_eswitch *esw, { int err, err1; - mlx5_eswitch_disable(esw, true); + mlx5_eswitch_disable(esw, false); err = mlx5_eswitch_enable(esw, MLX5_ESWITCH_LEGACY); if (err) { NL_SET_ERR_MSG_MOD(extack, "Failed setting eswitch to legacy"); From 76781623f009d5615b67f0675230ef90eaa9272a Mon Sep 17 00:00:00 2001 From: Paul Blakey Date: Tue, 31 Dec 2019 17:04:15 +0200 Subject: [PATCH 6/7] net/mlx5: Fix lowest FDB pool size The pool sizes represent the pool sizes in the fw. when we request a pool size from fw, it will return the next possible group. We track how many pools the fw has left and start requesting groups from the big to the small. When we start request 4k group, which doesn't exists in fw, fw wants to allocate the next possible size, 64k, but will fail since its exhausted. The correct smallest pool size in fw is 128 and not 4k. Fixes: 39ac237ce009 ("net/mlx5: E-Switch, Refactor chains and priorities") Signed-off-by: Paul Blakey Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.c index c5a446e295aa..4276194b633f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.c @@ -35,7 +35,7 @@ static const unsigned int ESW_POOLS[] = { 4 * 1024 * 1024, 1 * 1024 * 1024, 64 * 1024, - 4 * 1024, }; + 128 }; struct mlx5_esw_chains_priv { struct rhashtable chains_ht; From 13a7e459a41a56d788ab33d825c6205379bbb711 Mon Sep 17 00:00:00 2001 From: Erez Shitrit Date: Tue, 14 Jan 2020 09:27:27 +0200 Subject: [PATCH 7/7] net/mlx5: DR, Handle reformat capability over sw-steering tables On flow table creation, send the relevant flags according to what the FW currently supports. When FW doesn't support reformat option over SW-steering managed table, the driver shouldn't pass this. Fixes: 988fd6b32d07 ("net/mlx5: DR, Pass table flags at creation to lower layer") Signed-off-by: Erez Shitrit Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c | 9 +++++++-- include/linux/mlx5/mlx5_ifc.h | 5 ++++- 2 files changed, 11 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c index 3abfc8125926..c2027192e21e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c @@ -66,15 +66,20 @@ static int mlx5_cmd_dr_create_flow_table(struct mlx5_flow_root_namespace *ns, struct mlx5_flow_table *next_ft) { struct mlx5dr_table *tbl; + u32 flags; int err; if (mlx5_dr_is_fw_table(ft->flags)) return mlx5_fs_cmd_get_fw_cmds()->create_flow_table(ns, ft, log_size, next_ft); + flags = ft->flags; + /* turn off encap/decap if not supported for sw-str by fw */ + if (!MLX5_CAP_FLOWTABLE(ns->dev, sw_owner_reformat_supported)) + flags = ft->flags & ~(MLX5_FLOW_TABLE_TUNNEL_EN_REFORMAT | + MLX5_FLOW_TABLE_TUNNEL_EN_DECAP); - tbl = mlx5dr_table_create(ns->fs_dr_domain.dr_domain, - ft->level, ft->flags); + tbl = mlx5dr_table_create(ns->fs_dr_domain.dr_domain, ft->level, flags); if (!tbl) { mlx5_core_err(ns->dev, "Failed creating dr flow_table\n"); return -EINVAL; diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index ff8c9d527bb4..bfdf41537cf1 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -688,7 +688,10 @@ struct mlx5_ifc_flow_table_nic_cap_bits { u8 nic_rx_multi_path_tirs[0x1]; u8 nic_rx_multi_path_tirs_fts[0x1]; u8 allow_sniffer_and_nic_rx_shared_tir[0x1]; - u8 reserved_at_3[0x1d]; + u8 reserved_at_3[0x4]; + u8 sw_owner_reformat_supported[0x1]; + u8 reserved_at_8[0x18]; + u8 encap_general_header[0x1]; u8 reserved_at_21[0xa]; u8 log_max_packet_reformat_context[0x5];