From 2700b7e603af39ca55fe9fc876ca123efd44680f Mon Sep 17 00:00:00 2001 From: Shay Drory Date: Tue, 24 Feb 2026 13:46:48 +0200 Subject: net/mlx5: DR, Fix circular locking dependency in dump Fix a circular locking dependency between dbg_mutex and the domain rx/tx mutexes that could lead to a deadlock. The dump path in dr_dump_domain_all() was acquiring locks in the order: dbg_mutex -> rx.mutex -> tx.mutex While the table/matcher creation paths acquire locks in the order: rx.mutex -> tx.mutex -> dbg_mutex This inverted lock ordering creates a circular dependency. Fix this by changing dr_dump_domain_all() to acquire the domain lock before dbg_mutex, matching the order used in mlx5dr_table_create() and mlx5dr_matcher_create(). Lockdep splat: ====================================================== WARNING: possible circular locking dependency detected 6.19.0-rc6net_next_e817c4e #1 Not tainted ------------------------------------------------------ sos/30721 is trying to acquire lock: ffff888102df5900 (&dmn->info.rx.mutex){+.+.}-{4:4}, at: dr_dump_start+0x131/0x450 [mlx5_core] but task is already holding lock: ffff888102df5bc0 (&dmn->dump_info.dbg_mutex){+.+.}-{4:4}, at: dr_dump_start+0x10b/0x450 [mlx5_core] which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #2 (&dmn->dump_info.dbg_mutex){+.+.}-{4:4}: __mutex_lock+0x91/0x1060 mlx5dr_matcher_create+0x377/0x5e0 [mlx5_core] mlx5_cmd_dr_create_flow_group+0x62/0xd0 [mlx5_core] mlx5_create_flow_group+0x113/0x1c0 [mlx5_core] mlx5_chains_create_prio+0x453/0x2290 [mlx5_core] mlx5_chains_get_table+0x2e2/0x980 [mlx5_core] esw_chains_create+0x1e6/0x3b0 [mlx5_core] esw_create_offloads_fdb_tables.cold+0x62/0x63f [mlx5_core] esw_offloads_enable+0x76f/0xd20 [mlx5_core] mlx5_eswitch_enable_locked+0x35a/0x500 [mlx5_core] mlx5_devlink_eswitch_mode_set+0x561/0x950 [mlx5_core] devlink_nl_eswitch_set_doit+0x67/0xe0 genl_family_rcv_msg_doit+0xe0/0x130 genl_rcv_msg+0x188/0x290 netlink_rcv_skb+0x4b/0xf0 genl_rcv+0x24/0x40 netlink_unicast+0x1ed/0x2c0 netlink_sendmsg+0x210/0x450 __sock_sendmsg+0x38/0x60 __sys_sendto+0x119/0x180 __x64_sys_sendto+0x20/0x30 do_syscall_64+0x70/0xd00 entry_SYSCALL_64_after_hwframe+0x4b/0x53 -> #1 (&dmn->info.tx.mutex){+.+.}-{4:4}: __mutex_lock+0x91/0x1060 mlx5dr_table_create+0x11d/0x530 [mlx5_core] mlx5_cmd_dr_create_flow_table+0x62/0x140 [mlx5_core] __mlx5_create_flow_table+0x46f/0x960 [mlx5_core] mlx5_create_flow_table+0x16/0x20 [mlx5_core] esw_create_offloads_fdb_tables+0x136/0x240 [mlx5_core] esw_offloads_enable+0x76f/0xd20 [mlx5_core] mlx5_eswitch_enable_locked+0x35a/0x500 [mlx5_core] mlx5_devlink_eswitch_mode_set+0x561/0x950 [mlx5_core] devlink_nl_eswitch_set_doit+0x67/0xe0 genl_family_rcv_msg_doit+0xe0/0x130 genl_rcv_msg+0x188/0x290 netlink_rcv_skb+0x4b/0xf0 genl_rcv+0x24/0x40 netlink_unicast+0x1ed/0x2c0 netlink_sendmsg+0x210/0x450 __sock_sendmsg+0x38/0x60 __sys_sendto+0x119/0x180 __x64_sys_sendto+0x20/0x30 do_syscall_64+0x70/0xd00 entry_SYSCALL_64_after_hwframe+0x4b/0x53 -> #0 (&dmn->info.rx.mutex){+.+.}-{4:4}: __lock_acquire+0x18b6/0x2eb0 lock_acquire+0xd3/0x2c0 __mutex_lock+0x91/0x1060 dr_dump_start+0x131/0x450 [mlx5_core] seq_read_iter+0xe3/0x410 seq_read+0xfb/0x130 full_proxy_read+0x53/0x80 vfs_read+0xba/0x330 ksys_read+0x65/0xe0 do_syscall_64+0x70/0xd00 entry_SYSCALL_64_after_hwframe+0x4b/0x53 Possible unsafe locking scenario: CPU0 CPU1 ---- ---- lock(&dmn->dump_info.dbg_mutex); lock(&dmn->info.tx.mutex); lock(&dmn->dump_info.dbg_mutex); lock(&dmn->info.rx.mutex); *** DEADLOCK *** Fixes: 9222f0b27da2 ("net/mlx5: DR, Add support for dumping steering info") Signed-off-by: Shay Drory Reviewed-by: Yevgeny Kliteynik Reviewed-by: Alex Vesker Signed-off-by: Tariq Toukan Reviewed-by: Simon Horman Link: https://patch.msgid.link/20260224114652.1787431-2-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlx5/core/steering/sws/dr_dbg.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/sws/dr_dbg.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/sws/dr_dbg.c index 8803fa071c50..18362e9c3314 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/sws/dr_dbg.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/sws/dr_dbg.c @@ -1051,8 +1051,8 @@ static int dr_dump_domain_all(struct seq_file *file, struct mlx5dr_domain *dmn) struct mlx5dr_table *tbl; int ret; - mutex_lock(&dmn->dump_info.dbg_mutex); mlx5dr_domain_lock(dmn); + mutex_lock(&dmn->dump_info.dbg_mutex); ret = dr_dump_domain(file, dmn); if (ret < 0) @@ -1065,8 +1065,8 @@ static int dr_dump_domain_all(struct seq_file *file, struct mlx5dr_domain *dmn) } unlock_mutex: - mlx5dr_domain_unlock(dmn); mutex_unlock(&dmn->dump_info.dbg_mutex); + mlx5dr_domain_unlock(dmn); return ret; } -- cgit v1.2.3 From bd7b9f83fb9f85228c3ac9748d9cba9fab7fb5a2 Mon Sep 17 00:00:00 2001 From: Shay Drory Date: Tue, 24 Feb 2026 13:46:49 +0200 Subject: net/mlx5: LAG, disable MPESW in lag_disable_change() mlx5_lag_disable_change() unconditionally called mlx5_disable_lag() when LAG was active, which is incorrect for MLX5_LAG_MODE_MPESW. Hnece, call mlx5_disable_mpesw() when running in MPESW mode. Fixes: a32327a3a02c ("net/mlx5: Lag, Control MultiPort E-Switch single FDB mode") Signed-off-by: Shay Drory Reviewed-by: Mark Bloch Signed-off-by: Tariq Toukan Reviewed-by: Simon Horman Link: https://patch.msgid.link/20260224114652.1787431-3-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c | 8 ++++++-- drivers/net/ethernet/mellanox/mlx5/core/lag/mpesw.c | 8 ++++---- drivers/net/ethernet/mellanox/mlx5/core/lag/mpesw.h | 5 +++++ 3 files changed, 15 insertions(+), 6 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c index 9fe47c836ebd..859f042caf79 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c @@ -1869,8 +1869,12 @@ void mlx5_lag_disable_change(struct mlx5_core_dev *dev) mutex_lock(&ldev->lock); ldev->mode_changes_in_progress++; - if (__mlx5_lag_is_active(ldev)) - mlx5_disable_lag(ldev); + if (__mlx5_lag_is_active(ldev)) { + if (ldev->mode == MLX5_LAG_MODE_MPESW) + mlx5_lag_disable_mpesw(ldev); + else + mlx5_disable_lag(ldev); + } mutex_unlock(&ldev->lock); mlx5_devcom_comp_unlock(dev->priv.hca_devcom_comp); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag/mpesw.c b/drivers/net/ethernet/mellanox/mlx5/core/lag/mpesw.c index 04762562d7d9..a63d48d18878 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag/mpesw.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag/mpesw.c @@ -65,7 +65,7 @@ err_metadata: return err; } -static int enable_mpesw(struct mlx5_lag *ldev) +static int mlx5_lag_enable_mpesw(struct mlx5_lag *ldev) { struct mlx5_core_dev *dev0; int err; @@ -126,7 +126,7 @@ err_add_devices: return err; } -static void disable_mpesw(struct mlx5_lag *ldev) +void mlx5_lag_disable_mpesw(struct mlx5_lag *ldev) { if (ldev->mode == MLX5_LAG_MODE_MPESW) { mlx5_mpesw_metadata_cleanup(ldev); @@ -152,9 +152,9 @@ static void mlx5_mpesw_work(struct work_struct *work) } if (mpesww->op == MLX5_MPESW_OP_ENABLE) - mpesww->result = enable_mpesw(ldev); + mpesww->result = mlx5_lag_enable_mpesw(ldev); else if (mpesww->op == MLX5_MPESW_OP_DISABLE) - disable_mpesw(ldev); + mlx5_lag_disable_mpesw(ldev); unlock: mutex_unlock(&ldev->lock); mlx5_devcom_comp_unlock(devcom); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag/mpesw.h b/drivers/net/ethernet/mellanox/mlx5/core/lag/mpesw.h index f5d9b5c97b0d..b767dbb4f457 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag/mpesw.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag/mpesw.h @@ -31,6 +31,11 @@ int mlx5_lag_mpesw_do_mirred(struct mlx5_core_dev *mdev, bool mlx5_lag_is_mpesw(struct mlx5_core_dev *dev); void mlx5_lag_mpesw_disable(struct mlx5_core_dev *dev); int mlx5_lag_mpesw_enable(struct mlx5_core_dev *dev); +#ifdef CONFIG_MLX5_ESWITCH +void mlx5_lag_disable_mpesw(struct mlx5_lag *ldev); +#else +static inline void mlx5_lag_disable_mpesw(struct mlx5_lag *ldev) {} +#endif /* CONFIG_MLX5_ESWITCH */ #ifdef CONFIG_MLX5_ESWITCH void mlx5_mpesw_speed_update_work(struct work_struct *work); -- cgit v1.2.3 From d7073e8b978ae925f1f0f08754f33f84d8547ea7 Mon Sep 17 00:00:00 2001 From: Shay Drory Date: Tue, 24 Feb 2026 13:46:50 +0200 Subject: net/mlx5: E-switch, Clear legacy flag when moving to switchdev The cited commit introduced MLX5_PRIV_FLAGS_SWITCH_LEGACY to identify when a transition to legacy mode is requested via devlink. However, the logic failed to clear this flag if the mode was subsequently changed back to MLX5_ESWITCH_OFFLOADS (switchdev). Consequently, if a user toggled from legacy to switchdev, the flag remained set, leaving the driver with wrong state indicating Fix this by explicitly clearing the MLX5_PRIV_FLAGS_SWITCH_LEGACY bit when the requested mode is MLX5_ESWITCH_OFFLOADS. Fixes: 2a4f56fbcc47 ("net/mlx5e: Keep netdev when leave switchdev for devlink set legacy only") Signed-off-by: Shay Drory Reviewed-by: Mark Bloch Signed-off-by: Tariq Toukan Reviewed-by: Simon Horman Link: https://patch.msgid.link/20260224114652.1787431-4-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 1b439cef3719..9d51d030596c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -4068,6 +4068,8 @@ int mlx5_devlink_eswitch_mode_set(struct devlink *devlink, u16 mode, if (mlx5_mode == MLX5_ESWITCH_LEGACY) esw->dev->priv.flags |= MLX5_PRIV_FLAGS_SWITCH_LEGACY; + if (mlx5_mode == MLX5_ESWITCH_OFFLOADS) + esw->dev->priv.flags &= ~MLX5_PRIV_FLAGS_SWITCH_LEGACY; mlx5_eswitch_disable_locked(esw); if (mlx5_mode == MLX5_ESWITCH_OFFLOADS) { if (mlx5_devlink_trap_get_num_active(esw->dev)) { -- cgit v1.2.3 From 60253042c0b87b61596368489c44d12ba720d11c Mon Sep 17 00:00:00 2001 From: Shay Drory Date: Tue, 24 Feb 2026 13:46:51 +0200 Subject: net/mlx5: Fix missing devlink lock in SRIOV enable error path The cited commit miss to add locking in the error path of mlx5_sriov_enable(). When pci_enable_sriov() fails, mlx5_device_disable_sriov() is called to clean up. This cleanup function now expects to be called with the devlink instance lock held. Add the missing devl_lock(devlink) and devl_unlock(devlink) Fixes: 84a433a40d0e ("net/mlx5: Lock mlx5 devlink reload callbacks") Signed-off-by: Shay Drory Reviewed-by: Mark Bloch Signed-off-by: Tariq Toukan Reviewed-by: Simon Horman Link: https://patch.msgid.link/20260224114652.1787431-5-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlx5/core/sriov.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/sriov.c b/drivers/net/ethernet/mellanox/mlx5/core/sriov.c index a2fc937d5461..172862a70c70 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/sriov.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/sriov.c @@ -193,7 +193,9 @@ static int mlx5_sriov_enable(struct pci_dev *pdev, int num_vfs) err = pci_enable_sriov(pdev, num_vfs); if (err) { mlx5_core_warn(dev, "pci_enable_sriov failed : %d\n", err); + devl_lock(devlink); mlx5_device_disable_sriov(dev, num_vfs, true, true); + devl_unlock(devlink); } return err; } -- cgit v1.2.3 From 859380694f434597407632c29f30fdb5e763e6cc Mon Sep 17 00:00:00 2001 From: Jianbo Liu Date: Tue, 24 Feb 2026 13:46:52 +0200 Subject: net/mlx5e: Fix "scheduling while atomic" in IPsec MAC address query Fix a "scheduling while atomic" bug in mlx5e_ipsec_init_macs() by replacing mlx5_query_mac_address() with ether_addr_copy() to get the local MAC address directly from netdev->dev_addr. The issue occurs because mlx5_query_mac_address() queries the hardware which involves mlx5_cmd_exec() that can sleep, but it is called from the mlx5e_ipsec_handle_event workqueue which runs in atomic context. The MAC address is already available in netdev->dev_addr, so no need to query hardware. This avoids the sleeping call and resolves the bug. Call trace: BUG: scheduling while atomic: kworker/u112:2/69344/0x00000200 __schedule+0x7ab/0xa20 schedule+0x1c/0xb0 schedule_timeout+0x6e/0xf0 __wait_for_common+0x91/0x1b0 cmd_exec+0xa85/0xff0 [mlx5_core] mlx5_cmd_exec+0x1f/0x50 [mlx5_core] mlx5_query_nic_vport_mac_address+0x7b/0xd0 [mlx5_core] mlx5_query_mac_address+0x19/0x30 [mlx5_core] mlx5e_ipsec_init_macs+0xc1/0x720 [mlx5_core] mlx5e_ipsec_build_accel_xfrm_attrs+0x422/0x670 [mlx5_core] mlx5e_ipsec_handle_event+0x2b9/0x460 [mlx5_core] process_one_work+0x178/0x2e0 worker_thread+0x2ea/0x430 Fixes: cee137a63431 ("net/mlx5e: Handle ESN update events") Signed-off-by: Jianbo Liu Reviewed-by: Leon Romanovsky Signed-off-by: Tariq Toukan Reviewed-by: Simon Horman Link: https://patch.msgid.link/20260224114652.1787431-6-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c index 9c7064187ed0..f03507a522b4 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c @@ -259,7 +259,6 @@ static void mlx5e_ipsec_init_limits(struct mlx5e_ipsec_sa_entry *sa_entry, static void mlx5e_ipsec_init_macs(struct mlx5e_ipsec_sa_entry *sa_entry, struct mlx5_accel_esp_xfrm_attrs *attrs) { - struct mlx5_core_dev *mdev = mlx5e_ipsec_sa2dev(sa_entry); struct mlx5e_ipsec_addr *addrs = &attrs->addrs; struct net_device *netdev = sa_entry->dev; struct xfrm_state *x = sa_entry->x; @@ -276,7 +275,7 @@ static void mlx5e_ipsec_init_macs(struct mlx5e_ipsec_sa_entry *sa_entry, attrs->type != XFRM_DEV_OFFLOAD_PACKET) return; - mlx5_query_mac_address(mdev, addr); + ether_addr_copy(addr, netdev->dev_addr); switch (attrs->dir) { case XFRM_DEV_OFFLOAD_IN: src = attrs->dmac; -- cgit v1.2.3