From 1e66220948df815d7b37e0ff8b4627ce10433738 Mon Sep 17 00:00:00 2001 From: Adham Faris Date: Sun, 8 Jan 2023 18:09:32 +0200 Subject: [PATCH 01/10] net/mlx5e: Update rx ring hw mtu upon each rx-fcs flag change rq->hw_mtu is used in function en_rx.c/mlx5e_skb_from_cqe_mpwrq_linear() to catch oversized packets. If FCS is concatenated to the end of the packet then the check should be updated accordingly. Rx rings initialization (mlx5e_init_rxq_rq()) invoked for every new set of channels, as part of mlx5e_safe_switch_params(), unknowingly if it runs with default configuration or not. Current rq->hw_mtu initialization assumes default configuration and ignores params->scatter_fcs_en flag state. Fix this, by accounting for params->scatter_fcs_en flag state during rq->hw_mtu initialization. In addition, updating rq->hw_mtu value during ingress traffic might lead to packets drop and oversize_pkts_sw_drop counter increase with no good reason. Hence we remove this optimization and switch the set of channels with a new one, to make sure we don't get false positives on the oversize_pkts_sw_drop counter. Fixes: 102722fc6832 ("net/mlx5e: Add support for RXFCS feature flag") Signed-off-by: Adham Faris Reviewed-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/en_main.c | 86 ++++--------------- 1 file changed, 15 insertions(+), 71 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index abcc614b6191..f8b5e5993f35 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -591,7 +591,8 @@ static int mlx5e_init_rxq_rq(struct mlx5e_channel *c, struct mlx5e_params *param rq->ix = c->ix; rq->channel = c; rq->mdev = mdev; - rq->hw_mtu = MLX5E_SW2HW_MTU(params, params->sw_mtu); + rq->hw_mtu = + MLX5E_SW2HW_MTU(params, params->sw_mtu) - ETH_FCS_LEN * !params->scatter_fcs_en; rq->xdpsq = &c->rq_xdpsq; rq->stats = &c->priv->channel_stats[c->ix]->rq; rq->ptp_cyc2time = mlx5_rq_ts_translator(mdev); @@ -1014,35 +1015,6 @@ int mlx5e_flush_rq(struct mlx5e_rq *rq, int curr_state) return mlx5e_rq_to_ready(rq, curr_state); } -static int mlx5e_modify_rq_scatter_fcs(struct mlx5e_rq *rq, bool enable) -{ - struct mlx5_core_dev *mdev = rq->mdev; - - void *in; - void *rqc; - int inlen; - int err; - - inlen = MLX5_ST_SZ_BYTES(modify_rq_in); - in = kvzalloc(inlen, GFP_KERNEL); - if (!in) - return -ENOMEM; - - rqc = MLX5_ADDR_OF(modify_rq_in, in, ctx); - - MLX5_SET(modify_rq_in, in, rq_state, MLX5_RQC_STATE_RDY); - MLX5_SET64(modify_rq_in, in, modify_bitmask, - MLX5_MODIFY_RQ_IN_MODIFY_BITMASK_SCATTER_FCS); - MLX5_SET(rqc, rqc, scatter_fcs, enable); - MLX5_SET(rqc, rqc, state, MLX5_RQC_STATE_RDY); - - err = mlx5_core_modify_rq(mdev, rq->rqn, in); - - kvfree(in); - - return err; -} - static int mlx5e_modify_rq_vsd(struct mlx5e_rq *rq, bool vsd) { struct mlx5_core_dev *mdev = rq->mdev; @@ -3314,20 +3286,6 @@ static void mlx5e_cleanup_nic_tx(struct mlx5e_priv *priv) mlx5e_destroy_tises(priv); } -static int mlx5e_modify_channels_scatter_fcs(struct mlx5e_channels *chs, bool enable) -{ - int err = 0; - int i; - - for (i = 0; i < chs->num; i++) { - err = mlx5e_modify_rq_scatter_fcs(&chs->c[i]->rq, enable); - if (err) - return err; - } - - return 0; -} - static int mlx5e_modify_channels_vsd(struct mlx5e_channels *chs, bool vsd) { int err; @@ -3903,41 +3861,27 @@ static int mlx5e_set_rx_port_ts(struct mlx5_core_dev *mdev, bool enable) return mlx5_set_ports_check(mdev, in, sizeof(in)); } +static int mlx5e_set_rx_port_ts_wrap(struct mlx5e_priv *priv, void *ctx) +{ + struct mlx5_core_dev *mdev = priv->mdev; + bool enable = *(bool *)ctx; + + return mlx5e_set_rx_port_ts(mdev, enable); +} + static int set_feature_rx_fcs(struct net_device *netdev, bool enable) { struct mlx5e_priv *priv = netdev_priv(netdev); struct mlx5e_channels *chs = &priv->channels; - struct mlx5_core_dev *mdev = priv->mdev; + struct mlx5e_params new_params; int err; mutex_lock(&priv->state_lock); - if (enable) { - err = mlx5e_set_rx_port_ts(mdev, false); - if (err) - goto out; - - chs->params.scatter_fcs_en = true; - err = mlx5e_modify_channels_scatter_fcs(chs, true); - if (err) { - chs->params.scatter_fcs_en = false; - mlx5e_set_rx_port_ts(mdev, true); - } - } else { - chs->params.scatter_fcs_en = false; - err = mlx5e_modify_channels_scatter_fcs(chs, false); - if (err) { - chs->params.scatter_fcs_en = true; - goto out; - } - err = mlx5e_set_rx_port_ts(mdev, true); - if (err) { - mlx5_core_warn(mdev, "Failed to set RX port timestamp %d\n", err); - err = 0; - } - } - -out: + new_params = chs->params; + new_params.scatter_fcs_en = enable; + err = mlx5e_safe_switch_params(priv, &new_params, mlx5e_set_rx_port_ts_wrap, + &new_params.scatter_fcs_en, true); mutex_unlock(&priv->state_lock); return err; } From 288d85e07fbca5ee35882a4884dd701e43637be1 Mon Sep 17 00:00:00 2001 From: Yevgeny Kliteynik Date: Tue, 17 Jan 2023 17:34:20 +0200 Subject: [PATCH 02/10] net/mlx5: DR, Fix potential race in dr_rule_create_rule_nic Selecting builder should be protected by the lock to prevent the case where a new rule sets a builder in the nic_matcher while the previous rule is still using the nic_matcher. Fixing this issue and cleaning the error flow. Fixes: b9b81e1e9382 ("net/mlx5: DR, For short chains of STEs, avoid allocating ste_arr dynamically") Signed-off-by: Yevgeny Kliteynik Reviewed-by: Alex Vesker Signed-off-by: Saeed Mahameed --- .../mellanox/mlx5/core/steering/dr_rule.c | 25 +++++++++++-------- 1 file changed, 15 insertions(+), 10 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_rule.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_rule.c index b851141e03de..042ca0349124 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_rule.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_rule.c @@ -1138,12 +1138,14 @@ dr_rule_create_rule_nic(struct mlx5dr_rule *rule, rule->flow_source)) return 0; + mlx5dr_domain_nic_lock(nic_dmn); + ret = mlx5dr_matcher_select_builders(matcher, nic_matcher, dr_rule_get_ipv(¶m->outer), dr_rule_get_ipv(¶m->inner)); if (ret) - return ret; + goto err_unlock; hw_ste_arr_is_opt = nic_matcher->num_of_builders <= DR_RULE_MAX_STES_OPTIMIZED; if (likely(hw_ste_arr_is_opt)) { @@ -1152,12 +1154,12 @@ dr_rule_create_rule_nic(struct mlx5dr_rule *rule, hw_ste_arr = kzalloc((nic_matcher->num_of_builders + DR_ACTION_MAX_STES) * DR_STE_SIZE, GFP_KERNEL); - if (!hw_ste_arr) - return -ENOMEM; + if (!hw_ste_arr) { + ret = -ENOMEM; + goto err_unlock; + } } - mlx5dr_domain_nic_lock(nic_dmn); - ret = mlx5dr_matcher_add_to_tbl_nic(dmn, nic_matcher); if (ret) goto free_hw_ste; @@ -1223,7 +1225,10 @@ dr_rule_create_rule_nic(struct mlx5dr_rule *rule, mlx5dr_domain_nic_unlock(nic_dmn); - goto out; + if (unlikely(!hw_ste_arr_is_opt)) + kfree(hw_ste_arr); + + return 0; free_rule: dr_rule_clean_rule_members(rule, nic_rule); @@ -1238,12 +1243,12 @@ remove_from_nic_tbl: mlx5dr_matcher_remove_from_tbl_nic(dmn, nic_matcher); free_hw_ste: - mlx5dr_domain_nic_unlock(nic_dmn); - -out: - if (unlikely(!hw_ste_arr_is_opt)) + if (!hw_ste_arr_is_opt) kfree(hw_ste_arr); +err_unlock: + mlx5dr_domain_nic_unlock(nic_dmn); + return ret; } From da0c52426cd23f8728eff72c2b2d2a3eb6b451f5 Mon Sep 17 00:00:00 2001 From: Vlad Buslov Date: Thu, 26 Jan 2023 14:47:12 +0100 Subject: [PATCH 03/10] net/mlx5: Bridge, fix ageing of peer FDB entries SWITCHDEV_FDB_ADD_TO_BRIDGE event handler that updates FDB entry 'lastuse' field is only executed for eswitch that owns the entry. However, if peer entry processed packets at least once it will have hardware counter 'used' value greater than entry 'lastuse' from that point on, which will cause FDB entry not being aged out. Process the event on all eswitch instances. Fixes: ff9b7521468b ("net/mlx5: Bridge, support LAG") Signed-off-by: Vlad Buslov Reviewed-by: Maor Dickman Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en/rep/bridge.c | 4 ---- drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c | 2 +- 2 files changed, 1 insertion(+), 5 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/bridge.c b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/bridge.c index 8099a21e674c..ce85b48d327d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/bridge.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/bridge.c @@ -438,10 +438,6 @@ static int mlx5_esw_bridge_switchdev_event(struct notifier_block *nb, switch (event) { case SWITCHDEV_FDB_ADD_TO_BRIDGE: - /* only handle the event on native eswtich of representor */ - if (!mlx5_esw_bridge_is_local(dev, rep, esw)) - break; - fdb_info = container_of(info, struct switchdev_notifier_fdb_info, info); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c b/drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c index b176648d1343..3cdcb0e0b20f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c @@ -1715,7 +1715,7 @@ void mlx5_esw_bridge_fdb_update_used(struct net_device *dev, u16 vport_num, u16 struct mlx5_esw_bridge *bridge; port = mlx5_esw_bridge_port_lookup(vport_num, esw_owner_vhca_id, br_offloads); - if (!port || port->flags & MLX5_ESW_BRIDGE_PORT_FLAG_PEER) + if (!port) return; bridge = port->bridge; From 8974aa9638df557f4642acef707af15648a03555 Mon Sep 17 00:00:00 2001 From: Amir Tzin Date: Sun, 8 Jan 2023 15:54:46 +0200 Subject: [PATCH 04/10] net/mlx5e: Fix crash unsetting rx-vlan-filter in switchdev mode Moving to switchdev mode with rx-vlan-filter on and then setting it off causes the kernel to crash since fs->vlan is freed during nic profile cleanup flow. RX VLAN filtering is not supported in switchdev mode so unset it when changing to switchdev and restore its value when switching back to legacy. trace: [] RIP: 0010:mlx5e_disable_cvlan_filter+0x43/0x70 [] set_feature_cvlan_filter+0x37/0x40 [mlx5_core] [] mlx5e_handle_feature+0x3a/0x60 [mlx5_core] [] mlx5e_set_features+0x6d/0x160 [mlx5_core] [] __netdev_update_features+0x288/0xa70 [] ethnl_set_features+0x309/0x380 [] ? __nla_parse+0x21/0x30 [] genl_family_rcv_msg_doit.isra.17+0x110/0x150 [] genl_rcv_msg+0x112/0x260 [] ? features_reply_size+0xe0/0xe0 [] ? genl_family_rcv_msg_doit.isra.17+0x150/0x150 [] netlink_rcv_skb+0x4e/0x100 [] genl_rcv+0x24/0x40 [] netlink_unicast+0x1ab/0x290 [] netlink_sendmsg+0x257/0x4f0 [] sock_sendmsg+0x5c/0x70 Fixes: cb67b832921c ("net/mlx5e: Introduce SRIOV VF representors") Signed-off-by: Amir Tzin Reviewed-by: Maor Dickman Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_fs.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 4 ++++ 2 files changed, 5 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c index 1892ccb889b3..7cd36f4ac3ef 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c @@ -443,7 +443,7 @@ void mlx5e_enable_cvlan_filter(struct mlx5e_flow_steering *fs, bool promisc) void mlx5e_disable_cvlan_filter(struct mlx5e_flow_steering *fs, bool promisc) { - if (fs->vlan->cvlan_filter_disabled) + if (!fs->vlan || fs->vlan->cvlan_filter_disabled) return; fs->vlan->cvlan_filter_disabled = true; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index f8b5e5993f35..6c24f33a5ea5 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -4018,6 +4018,10 @@ static netdev_features_t mlx5e_fix_uplink_rep_features(struct net_device *netdev if (netdev->features & NETIF_F_GRO_HW) netdev_warn(netdev, "Disabling HW_GRO, not supported in switchdev mode\n"); + features &= ~NETIF_F_HW_VLAN_CTAG_FILTER; + if (netdev->features & NETIF_F_HW_VLAN_CTAG_FILTER) + netdev_warn(netdev, "Disabling HW_VLAN CTAG FILTERING, not supported in switchdev mode\n"); + return features; } From 8aa5f171d51c1cb69e5e3106df4dd1a446102823 Mon Sep 17 00:00:00 2001 From: Dragos Tatulea Date: Mon, 23 Jan 2023 16:06:32 +0200 Subject: [PATCH 05/10] net/mlx5e: IPoIB, Show unknown speed instead of error ethtool is returning an error for unknown speeds for the IPoIB interface: $ ethtool ib0 netlink error: failed to retrieve link settings netlink error: Invalid argument netlink error: failed to retrieve link settings netlink error: Invalid argument Settings for ib0: Link detected: no After this change, ethtool will return success and show "unknown speed": $ ethtool ib0 Settings for ib0: Supported ports: [ ] Supported link modes: Not reported Supported pause frame use: No Supports auto-negotiation: No Supported FEC modes: Not reported Advertised link modes: Not reported Advertised pause frame use: No Advertised auto-negotiation: No Advertised FEC modes: Not reported Speed: Unknown! Duplex: Full Auto-negotiation: off Port: Other PHYAD: 0 Transceiver: internal Link detected: no Fixes: eb234ee9d541 ("net/mlx5e: IPoIB, Add support for get_link_ksettings in ethtool") Signed-off-by: Dragos Tatulea Reviewed-by: Gal Pressman Reviewed-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c | 13 +++++-------- 1 file changed, 5 insertions(+), 8 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c index eff92dc0927c..e09518f887a0 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c @@ -189,16 +189,16 @@ static inline int mlx5_ptys_rate_enum_to_int(enum mlx5_ptys_rate rate) } } -static int mlx5i_get_speed_settings(u16 ib_link_width_oper, u16 ib_proto_oper) +static u32 mlx5i_get_speed_settings(u16 ib_link_width_oper, u16 ib_proto_oper) { int rate, width; rate = mlx5_ptys_rate_enum_to_int(ib_proto_oper); if (rate < 0) - return -EINVAL; + return SPEED_UNKNOWN; width = mlx5_ptys_width_enum_to_int(ib_link_width_oper); if (width < 0) - return -EINVAL; + return SPEED_UNKNOWN; return rate * width; } @@ -221,16 +221,13 @@ static int mlx5i_get_link_ksettings(struct net_device *netdev, ethtool_link_ksettings_zero_link_mode(link_ksettings, advertising); speed = mlx5i_get_speed_settings(ib_link_width_oper, ib_proto_oper); - if (speed < 0) - return -EINVAL; + link_ksettings->base.speed = speed; + link_ksettings->base.duplex = speed == SPEED_UNKNOWN ? DUPLEX_UNKNOWN : DUPLEX_FULL; - link_ksettings->base.duplex = DUPLEX_FULL; link_ksettings->base.port = PORT_OTHER; link_ksettings->base.autoneg = AUTONEG_DISABLE; - link_ksettings->base.speed = speed; - return 0; } From c3bdbaea654d8df39112de33037106134a520dc7 Mon Sep 17 00:00:00 2001 From: Maher Sanalla Date: Sun, 22 Jan 2023 21:09:40 +0200 Subject: [PATCH 06/10] net/mlx5: Store page counters in a single array Currently, an independent page counter is used for tracking memory usage for each function type such as VF, PF and host PF (DPU). For better code-readibilty, use a single array that stores the number of allocated memory pages for each function type. Signed-off-by: Maher Sanalla Reviewed-by: Shay Drory Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/debugfs.c | 4 +- .../net/ethernet/mellanox/mlx5/core/ecpf.c | 2 +- .../ethernet/mellanox/mlx5/core/pagealloc.c | 37 +++++++++++-------- .../net/ethernet/mellanox/mlx5/core/sriov.c | 2 +- include/linux/mlx5/driver.h | 12 ++++-- 5 files changed, 34 insertions(+), 23 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c b/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c index 3e232a65a0c3..c3e7c24a0971 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c @@ -245,8 +245,8 @@ void mlx5_pages_debugfs_init(struct mlx5_core_dev *dev) pages = dev->priv.dbg.pages_debugfs; debugfs_create_u32("fw_pages_total", 0400, pages, &dev->priv.fw_pages); - debugfs_create_u32("fw_pages_vfs", 0400, pages, &dev->priv.vfs_pages); - debugfs_create_u32("fw_pages_host_pf", 0400, pages, &dev->priv.host_pf_pages); + debugfs_create_u32("fw_pages_vfs", 0400, pages, &dev->priv.page_counters[MLX5_VF]); + debugfs_create_u32("fw_pages_host_pf", 0400, pages, &dev->priv.page_counters[MLX5_HOST_PF]); debugfs_create_u32("fw_pages_alloc_failed", 0400, pages, &dev->priv.fw_pages_alloc_failed); debugfs_create_u32("fw_pages_give_dropped", 0400, pages, &dev->priv.give_pages_dropped); debugfs_create_u32("fw_pages_reclaim_discard", 0400, pages, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ecpf.c b/drivers/net/ethernet/mellanox/mlx5/core/ecpf.c index 464eb3a18450..cdc87ecae5d3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/ecpf.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/ecpf.c @@ -87,7 +87,7 @@ void mlx5_ec_cleanup(struct mlx5_core_dev *dev) mlx5_host_pf_cleanup(dev); - err = mlx5_wait_for_pages(dev, &dev->priv.host_pf_pages); + err = mlx5_wait_for_pages(dev, &dev->priv.page_counters[MLX5_HOST_PF]); if (err) mlx5_core_warn(dev, "Timeout reclaiming external host PF pages err(%d)\n", err); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c index 60596357bfc7..9f99292ab5ce 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c @@ -74,6 +74,14 @@ static u32 get_function(u16 func_id, bool ec_function) return (u32)func_id | (ec_function << 16); } +static u16 func_id_to_type(struct mlx5_core_dev *dev, u16 func_id, bool ec_function) +{ + if (!func_id) + return mlx5_core_is_ecpf(dev) && !ec_function ? MLX5_HOST_PF : MLX5_PF; + + return MLX5_VF; +} + static struct rb_root *page_root_per_function(struct mlx5_core_dev *dev, u32 function) { struct rb_root *root; @@ -332,6 +340,7 @@ static int give_pages(struct mlx5_core_dev *dev, u16 func_id, int npages, u32 out[MLX5_ST_SZ_DW(manage_pages_out)] = {0}; int inlen = MLX5_ST_SZ_BYTES(manage_pages_in); int notify_fail = event; + u16 func_type; u64 addr; int err; u32 *in; @@ -383,11 +392,9 @@ retry: goto out_dropped; } + func_type = func_id_to_type(dev, func_id, ec_function); + dev->priv.page_counters[func_type] += npages; dev->priv.fw_pages += npages; - if (func_id) - dev->priv.vfs_pages += npages; - else if (mlx5_core_is_ecpf(dev) && !ec_function) - dev->priv.host_pf_pages += npages; mlx5_core_dbg(dev, "npages %d, ec_function %d, func_id 0x%x, err %d\n", npages, ec_function, func_id, err); @@ -414,6 +421,7 @@ static void release_all_pages(struct mlx5_core_dev *dev, u16 func_id, struct rb_root *root; struct rb_node *p; int npages = 0; + u16 func_type; root = xa_load(&dev->priv.page_root_xa, function); if (WARN_ON_ONCE(!root)) @@ -428,11 +436,9 @@ static void release_all_pages(struct mlx5_core_dev *dev, u16 func_id, free_fwp(dev, fwp, fwp->free_count); } + func_type = func_id_to_type(dev, func_id, ec_function); + dev->priv.page_counters[func_type] -= npages; dev->priv.fw_pages -= npages; - if (func_id) - dev->priv.vfs_pages -= npages; - else if (mlx5_core_is_ecpf(dev) && !ec_function) - dev->priv.host_pf_pages -= npages; mlx5_core_dbg(dev, "npages %d, ec_function %d, func_id 0x%x\n", npages, ec_function, func_id); @@ -498,6 +504,7 @@ static int reclaim_pages(struct mlx5_core_dev *dev, u16 func_id, int npages, int outlen = MLX5_ST_SZ_BYTES(manage_pages_out); u32 in[MLX5_ST_SZ_DW(manage_pages_in)] = {}; int num_claimed; + u16 func_type; u32 *out; int err; int i; @@ -549,11 +556,9 @@ static int reclaim_pages(struct mlx5_core_dev *dev, u16 func_id, int npages, if (nclaimed) *nclaimed = num_claimed; + func_type = func_id_to_type(dev, func_id, ec_function); + dev->priv.page_counters[func_type] -= num_claimed; dev->priv.fw_pages -= num_claimed; - if (func_id) - dev->priv.vfs_pages -= num_claimed; - else if (mlx5_core_is_ecpf(dev) && !ec_function) - dev->priv.host_pf_pages -= num_claimed; out_free: kvfree(out); @@ -706,12 +711,12 @@ int mlx5_reclaim_startup_pages(struct mlx5_core_dev *dev) WARN(dev->priv.fw_pages, "FW pages counter is %d after reclaiming all pages\n", dev->priv.fw_pages); - WARN(dev->priv.vfs_pages, + WARN(dev->priv.page_counters[MLX5_VF], "VFs FW pages counter is %d after reclaiming all pages\n", - dev->priv.vfs_pages); - WARN(dev->priv.host_pf_pages, + dev->priv.page_counters[MLX5_VF]); + WARN(dev->priv.page_counters[MLX5_HOST_PF], "External host PF FW pages counter is %d after reclaiming all pages\n", - dev->priv.host_pf_pages); + dev->priv.page_counters[MLX5_HOST_PF]); return 0; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/sriov.c b/drivers/net/ethernet/mellanox/mlx5/core/sriov.c index c0e6c487c63c..3008e9ce2bbf 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/sriov.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/sriov.c @@ -147,7 +147,7 @@ mlx5_device_disable_sriov(struct mlx5_core_dev *dev, int num_vfs, bool clear_vf) mlx5_eswitch_disable_sriov(dev->priv.eswitch, clear_vf); - if (mlx5_wait_for_pages(dev, &dev->priv.vfs_pages)) + if (mlx5_wait_for_pages(dev, &dev->priv.page_counters[MLX5_VF])) mlx5_core_warn(dev, "timeout reclaiming VFs pages\n"); } diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 76ef2e4fde38..82a9bd4274b8 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -573,6 +573,13 @@ struct mlx5_debugfs_entries { struct dentry *lag_debugfs; }; +enum mlx5_func_type { + MLX5_PF, + MLX5_VF, + MLX5_HOST_PF, + MLX5_FUNC_TYPE_NUM, +}; + struct mlx5_ft_pool; struct mlx5_priv { /* IRQ table valid only for real pci devices PF or VF */ @@ -583,11 +590,10 @@ struct mlx5_priv { struct mlx5_nb pg_nb; struct workqueue_struct *pg_wq; struct xarray page_root_xa; - u32 fw_pages; atomic_t reg_pages; struct list_head free_list; - u32 vfs_pages; - u32 host_pf_pages; + u32 fw_pages; + u32 page_counters[MLX5_FUNC_TYPE_NUM]; u32 fw_pages_alloc_failed; u32 give_pages_dropped; u32 reclaim_pages_discard; From 9965bbebae59b3563a4d95e4aed121e8965dfdc2 Mon Sep 17 00:00:00 2001 From: Maher Sanalla Date: Sun, 22 Jan 2023 23:24:56 +0200 Subject: [PATCH 07/10] net/mlx5: Expose SF firmware pages counter Currently, each core device has VF pages counter which stores number of fw pages used by its VFs and SFs. The current design led to a hang when performing firmware reset on DPU, where the DPU PFs stalled in sriov unload flow due to waiting on release of SFs pages instead of waiting on only VFs pages. Thus, Add a separate counter for SF firmware pages, which will prevent the stall scenario described above. Fixes: 1958fc2f0712 ("net/mlx5: SF, Add auxiliary device driver") Signed-off-by: Maher Sanalla Reviewed-by: Shay Drory Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/debugfs.c | 1 + drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c | 2 +- include/linux/mlx5/driver.h | 1 + 3 files changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c b/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c index c3e7c24a0971..bb95b40d25eb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c @@ -246,6 +246,7 @@ void mlx5_pages_debugfs_init(struct mlx5_core_dev *dev) debugfs_create_u32("fw_pages_total", 0400, pages, &dev->priv.fw_pages); debugfs_create_u32("fw_pages_vfs", 0400, pages, &dev->priv.page_counters[MLX5_VF]); + debugfs_create_u32("fw_pages_sfs", 0400, pages, &dev->priv.page_counters[MLX5_SF]); debugfs_create_u32("fw_pages_host_pf", 0400, pages, &dev->priv.page_counters[MLX5_HOST_PF]); debugfs_create_u32("fw_pages_alloc_failed", 0400, pages, &dev->priv.fw_pages_alloc_failed); debugfs_create_u32("fw_pages_give_dropped", 0400, pages, &dev->priv.give_pages_dropped); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c index 9f99292ab5ce..0eb50be175cc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c @@ -79,7 +79,7 @@ static u16 func_id_to_type(struct mlx5_core_dev *dev, u16 func_id, bool ec_funct if (!func_id) return mlx5_core_is_ecpf(dev) && !ec_function ? MLX5_HOST_PF : MLX5_PF; - return MLX5_VF; + return func_id <= mlx5_core_max_vfs(dev) ? MLX5_VF : MLX5_SF; } static struct rb_root *page_root_per_function(struct mlx5_core_dev *dev, u32 function) diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 82a9bd4274b8..333c1fec72f8 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -576,6 +576,7 @@ struct mlx5_debugfs_entries { enum mlx5_func_type { MLX5_PF, MLX5_VF, + MLX5_SF, MLX5_HOST_PF, MLX5_FUNC_TYPE_NUM, }; From db561fed6b8fa3878e74d5df6512a4a38152b63e Mon Sep 17 00:00:00 2001 From: Shay Drory Date: Mon, 9 Jan 2023 15:27:40 +0200 Subject: [PATCH 08/10] net/mlx5: fw_tracer, Clear load bit when freeing string DBs buffers Whenever the driver is reading the string DBs into buffers, the driver is setting the load bit, but the driver never clears this bit. As a result, in case load bit is on and the driver query the device for new string DBs, the driver won't read again the string DBs. Fix it by clearing the load bit when query the device for new string DBs. Fixes: 2d69356752ff ("net/mlx5: Add support for fw live patch event") Signed-off-by: Shay Drory Reviewed-by: Moshe Shemesh Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c | 1 + 1 file changed, 1 insertion(+) 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 21831386b26e..d82e98a0cdfa 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c @@ -64,6 +64,7 @@ static int mlx5_query_mtrc_caps(struct mlx5_fw_tracer *tracer) MLX5_GET(mtrc_cap, out, num_string_trace); tracer->str_db.num_string_db = MLX5_GET(mtrc_cap, out, num_string_db); tracer->owner = !!MLX5_GET(mtrc_cap, out, trace_owner); + tracer->str_db.loaded = false; for (i = 0; i < tracer->str_db.num_string_db; i++) { mtrc_cap_sp = MLX5_ADDR_OF(mtrc_cap, out, string_db_param[i]); From 184e1e4474dbcfebc4dbd1fa823a329978f25506 Mon Sep 17 00:00:00 2001 From: Shay Drory Date: Wed, 25 Jan 2023 17:39:36 +0200 Subject: [PATCH 09/10] net/mlx5: fw_tracer, Zero consumer index when reloading the tracer When tracer is reloaded, the device will log the traces at the beginning of the log buffer. Also, driver is reading the log buffer in chunks in accordance to the consumer index. Hence, zero consumer index when reloading the tracer. Fixes: 4383cfcc65e7 ("net/mlx5: Add devlink reload") Signed-off-by: Shay Drory Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) 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 d82e98a0cdfa..5b05b884b5fb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c @@ -757,6 +757,7 @@ static int mlx5_fw_tracer_set_mtrc_conf(struct mlx5_fw_tracer *tracer) if (err) mlx5_core_warn(dev, "FWTracer: Failed to set tracer configurations %d\n", err); + tracer->buff.consumer_index = 0; return err; } @@ -821,7 +822,6 @@ static void mlx5_fw_tracer_ownership_change(struct work_struct *work) mlx5_core_dbg(tracer->dev, "FWTracer: ownership changed, current=(%d)\n", tracer->owner); if (tracer->owner) { tracer->owner = false; - tracer->buff.consumer_index = 0; return; } From 8f0d1451ecf7b3bd5a06ffc866c753d0f3ab4683 Mon Sep 17 00:00:00 2001 From: Shay Drory Date: Wed, 14 Dec 2022 22:16:23 +0200 Subject: [PATCH 10/10] net/mlx5: Serialize module cleanup with reload and remove Currently, remove and reload flows can run in parallel to module cleanup. This design is error prone. For example: aux_drivers callbacks are called from both cleanup and remove flows with different lockings, which can cause a deadlock[1]. Hence, serialize module cleanup with reload and remove. [1] cleanup remove ------- ------ auxiliary_driver_unregister(); devl_lock() auxiliary_device_delete(mlx5e_aux) device_lock(mlx5e_aux) devl_lock() device_lock(mlx5e_aux) Fixes: 912cebf420c2 ("net/mlx5e: Connect ethernet part to auxiliary bus") Signed-off-by: Shay Drory Reviewed-by: Moshe Shemesh Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/main.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 3d5f2a4b1fed..4e1b5757528a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -2110,7 +2110,7 @@ static int __init mlx5_init(void) mlx5_core_verify_params(); mlx5_register_debugfs(); - err = pci_register_driver(&mlx5_core_driver); + err = mlx5e_init(); if (err) goto err_debug; @@ -2118,16 +2118,16 @@ static int __init mlx5_init(void) if (err) goto err_sf; - err = mlx5e_init(); + err = pci_register_driver(&mlx5_core_driver); if (err) - goto err_en; + goto err_pci; return 0; -err_en: +err_pci: mlx5_sf_driver_unregister(); err_sf: - pci_unregister_driver(&mlx5_core_driver); + mlx5e_cleanup(); err_debug: mlx5_unregister_debugfs(); return err; @@ -2135,9 +2135,9 @@ err_debug: static void __exit mlx5_cleanup(void) { - mlx5e_cleanup(); - mlx5_sf_driver_unregister(); pci_unregister_driver(&mlx5_core_driver); + mlx5_sf_driver_unregister(); + mlx5e_cleanup(); mlx5_unregister_debugfs(); }