mlx5 misc fixes 2026-01-26

Hi,

This patchset provides misc bug fixes from the team to the mlx5 core and Eth drivers.

Thanks,
Tariq.

Gal Pressman (1):
net/mlx5e: Account for netdev stats in ndo_get_stats64

Mark Bloch (1):
net/mlx5e: TC, delete flows only for existing peers

Shay Drory (1):
net/mlx5: Fix Unbinding uplink-netdev in switchdev mode

 drivers/net/ethernet/mellanox/mlx5/core/dev.c | 14 ++++++++++
 .../net/ethernet/mellanox/mlx5/core/en_main.c | 21 ++++++++-------
 .../net/ethernet/mellanox/mlx5/core/en_tc.c   | 19 +++++++++-----
 .../net/ethernet/mellanox/mlx5/core/eswitch.h |  4 +++
 .../mellanox/mlx5/core/eswitch_offloads.c     | 26 +++++++++++++++++++
 .../ethernet/mellanox/mlx5/core/mlx5_core.h   |  1 +
 6 files changed, 70 insertions(+), 15 deletions(-)

base-commit: 709bbb015538dfd5c97308b77c950d41a4d95cd3

2.40.1

From: Shay Drory <[email protected]>

It is possible to unbind the uplink ETH driver while the E-Switch is in switchdev mode. This leads to netdevice reference counting issues[1], as the driver removal path was not designed to clean up from this state.

During uplink ETH driver removal (_mlx5e_remove), the code now waits for any concurrent E-Switch mode transition to finish. It then removes the REPs auxiliary device, if exists. This ensures a graceful cleanup.

[1]
unregister_netdevice: waiting for eth2 to become free. Usage count = 2
ref_tracker: netdev@00000000c912e04b has 1/1 users at

     ib_device_set_netdev+0x130/0x270 [ib_core]
     mlx5_ib_vport_rep_load+0xf4/0x3e0 [mlx5_ib]
     mlx5_esw_offloads_rep_load+0xc7/0xe0 [mlx5_core]
     esw_offloads_enable+0x583/0x900 [mlx5_core]
     mlx5_eswitch_enable_locked+0x1b2/0x290 [mlx5_core]
     mlx5_devlink_eswitch_mode_set+0x107/0x3e0 [mlx5_core]
     devlink_nl_eswitch_set_doit+0x60/0xd0
     genl_family_rcv_msg_doit+0xe0/0x130
     genl_rcv_msg+0x183/0x290
     netlink_rcv_skb+0x4b/0xf0
     genl_rcv+0x24/0x40
     netlink_unicast+0x255/0x380
     netlink_sendmsg+0x1f3/0x420
     __sock_sendmsg+0x38/0x60
     __sys_sendto+0x119/0x180
     __x64_sys_sendto+0x20/0x30

---
 drivers/net/ethernet/mellanox/mlx5/core/dev.c | 14 ++++++++++
 .../net/ethernet/mellanox/mlx5/core/en_main.c |  1 +
 .../net/ethernet/mellanox/mlx5/core/eswitch.h |  4 +++
 .../mellanox/mlx5/core/eswitch_offloads.c     | 26 +++++++++++++++++++
 .../ethernet/mellanox/mlx5/core/mlx5_core.h   |  1 +
 5 files changed, 46 insertions(+)

index 64c04f52990f..781e39b5aa1d 100644
@@ -575,3 +575,17 @@ bool mlx5_same_hw_devs(struct mlx5_core_dev *dev, struct mlx5_core_dev *peer_dev
 	return plen && flen && flen == plen &&
 		!memcmp(fsystem_guid, psystem_guid, flen);
 }
+
+void mlx5_core_reps_aux_devs_remove(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+
+	if (priv->adev[MLX5_INTERFACE_PROTOCOL_ETH])
+		device_lock_assert(&priv->adev[MLX5_INTERFACE_PROTOCOL_ETH]->adev.dev);
+	else
+		mlx5_core_err(dev, "ETH driver already removed\n");
+	if (priv->adev[MLX5_INTERFACE_PROTOCOL_IB_REP])
+		del_adev(&priv->adev[MLX5_INTERFACE_PROTOCOL_IB_REP]->adev);
+	if (priv->adev[MLX5_INTERFACE_PROTOCOL_ETH_REP])
+		del_adev(&priv->adev[MLX5_INTERFACE_PROTOCOL_ETH_REP]->adev);
+}

index 9042c8a388e4..f83359f7fdea 100644
@@ -6842,6 +6842,7 @@ static void _mlx5e_remove(struct auxiliary_device *adev)
 	struct mlx5e_priv *priv = netdev_priv(netdev);
 	struct mlx5_core_dev *mdev = edev->mdev;
 
+	mlx5_eswitch_safe_aux_devs_remove(mdev);
 	mlx5_core_uplink_netdev_set(mdev, NULL);
 
 	if (priv->profile)

index e7fe43799b23..714ad28e8445 100644
@@ -929,6 +929,7 @@ int mlx5_esw_ipsec_vf_packet_offload_set(struct mlx5_eswitch *esw, struct mlx5_v
 int mlx5_esw_ipsec_vf_packet_offload_supported(struct mlx5_core_dev *dev,
 					       u16 vport_num);
 bool mlx5_esw_host_functions_enabled(const struct mlx5_core_dev *dev);
+void mlx5_eswitch_safe_aux_devs_remove(struct mlx5_core_dev *dev);
 #else  /* CONFIG_MLX5_ESWITCH */
 /* eswitch API stubs */
 static inline int  mlx5_eswitch_init(struct mlx5_core_dev *dev) { return 0; }

@@ -1012,6 +1013,9 @@ mlx5_esw_vport_vhca_id(struct mlx5_eswitch *esw, u16 vportn, u16 *vhca_id)
 	return false;
 }
 
+static inline void
+mlx5_eswitch_safe_aux_devs_remove(struct mlx5_core_dev *dev) {}
+
 #endif /* CONFIG_MLX5_ESWITCH */
 
 #endif /* __MLX5_ESWITCH_H__ */

index ea94a727633f..02b7e474586d 100644
@@ -3981,6 +3981,32 @@ static bool mlx5_devlink_switchdev_active_mode_change(struct mlx5_eswitch *esw,
 	return true;
 }
 
+#define MLX5_ESW_HOLD_TIMEOUT_MS 7000
+#define MLX5_ESW_HOLD_RETRY_DELAY_MS 500
+
+void mlx5_eswitch_safe_aux_devs_remove(struct mlx5_core_dev *dev)
+{
+	unsigned long timeout;
+	bool hold_esw = true;
+
+	/* Wait for any concurrent eswitch mode transition to complete. */
+	if (!mlx5_esw_hold(dev)) {
+		timeout = jiffies + msecs_to_jiffies(MLX5_ESW_HOLD_TIMEOUT_MS);
+		while (!mlx5_esw_hold(dev)) {
+			if (!time_before(jiffies, timeout)) {
+				hold_esw = false;
+				break;
+			}
+			msleep(MLX5_ESW_HOLD_RETRY_DELAY_MS);
+		}
+	}
+	if (hold_esw) {
+		if (mlx5_eswitch_mode(dev) == MLX5_ESWITCH_OFFLOADS)
+			mlx5_core_reps_aux_devs_remove(dev);
+		mlx5_esw_release(dev);
+	}
+}
+
 int mlx5_devlink_eswitch_mode_set(struct devlink *devlink, u16 mode,
 				  struct netlink_ext_ack *extack)
 {

index cfebc110c02f..99b0a25054ef 100644
@@ -290,6 +290,7 @@ int mlx5_register_device(struct mlx5_core_dev *dev);
 void mlx5_unregister_device(struct mlx5_core_dev *dev);
 void mlx5_dev_set_lightweight(struct mlx5_core_dev *dev);
 bool mlx5_dev_is_lightweight(struct mlx5_core_dev *dev);
+void mlx5_core_reps_aux_devs_remove(struct mlx5_core_dev *dev);
 
 void mlx5_fw_reporters_create(struct mlx5_core_dev *dev);
 int mlx5_query_mtpps(struct mlx5_core_dev *dev, u32 *mtpps, u32 mtpps_size);

2.40.1

From: Mark Bloch <[email protected]>

When deleting TC steering flows, iterate only over actual devcom peers instead of assuming all possible ports exist. This avoids touching non-existent peers and ensures cleanup is limited to devices the driver is currently connected to.

 BUG: kernel NULL pointer dereference, address: 0000000000000008
 #PF: supervisor write access in kernel mode
 #PF: error_code(0x0002) - not-present page
 PGD 133c8a067 P4D 0
 Oops: Oops: 0002 [#1] SMP
 CPU: 19 UID: 0 PID: 2169 Comm: tc Not tainted 6.18.0+ #156 NONE
 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
 RIP: 0010:mlx5e_tc_del_fdb_peers_flow+0xbe/0x200 [mlx5_core]
 Code: 00 00 a8 08 74 a8 49 8b 46 18 f6 c4 02 74 9f 4c 8d bf a0 12 00 00 4c 89 ff e8 0e e7 96 e1 49 8b 44 24 08 49 8b 0c 24 4c 89 ff <48> 89 41 08 48 89 08 49 89 2c 24 49 89 5c 24 08 e8 7d ce 96 e1 49
 RSP: 0018:ff11000143867528 EFLAGS: 00010246
 RAX: 0000000000000000 RBX: dead000000000122 RCX: 0000000000000000
 RDX: ff11000143691580 RSI: ff110001026e5000 RDI: ff11000106f3d2a0
 RBP: dead000000000100 R08: 00000000000003fd R09: 0000000000000002
 R10: ff11000101c75690 R11: ff1100085faea178 R12: ff11000115f0ae78
 R13: 0000000000000000 R14: ff11000115f0a800 R15: ff11000106f3d2a0
 FS:  00007f35236bf740(0000) GS:ff110008dc809000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 0000000000000008 CR3: 0000000157a01001 CR4: 0000000000373eb0
 Call Trace:
  <TASK>
  mlx5e_tc_del_flow+0x46/0x270 [mlx5_core]
  mlx5e_flow_put+0x25/0x50 [mlx5_core]
  mlx5e_delete_flower+0x2a6/0x3e0 [mlx5_core]
  tc_setup_cb_reoffload+0x20/0x80
  fl_reoffload+0x26f/0x2f0 [cls_flower]
  ? mlx5e_tc_reoffload_flows_work+0xc0/0xc0 [mlx5_core]
  ? mlx5e_tc_reoffload_flows_work+0xc0/0xc0 [mlx5_core]
  tcf_block_playback_offloads+0x9e/0x1c0
  tcf_block_unbind+0x7b/0xd0
  tcf_block_setup+0x186/0x1d0
  tcf_block_offload_cmd.isra.0+0xef/0x130
  tcf_block_offload_unbind+0x43/0x70
  __tcf_block_put+0x85/0x160
  ingress_destroy+0x32/0x110 [sch_ingress]
  __qdisc_destroy+0x44/0x100
  qdisc_graft+0x22b/0x610
  tc_get_qdisc+0x183/0x4d0
  rtnetlink_rcv_msg+0x2d7/0x3d0
  ? rtnl_calcit.isra.0+0x100/0x100
  netlink_rcv_skb+0x53/0x100
  netlink_unicast+0x249/0x320
  ? __alloc_skb+0x102/0x1f0
  netlink_sendmsg+0x1e3/0x420
  __sock_sendmsg+0x38/0x60
  ____sys_sendmsg+0x1ef/0x230
  ? copy_msghdr_from_user+0x6c/0xa0
  ___sys_sendmsg+0x7f/0xc0
  ? ___sys_recvmsg+0x8a/0xc0
  ? __sys_sendto+0x119/0x180
  __sys_sendmsg+0x61/0xb0
  do_syscall_64+0x55/0x640
  entry_SYSCALL_64_after_hwframe+0x4b/0x53
 RIP: 0033:0x7f35238bb764
 Code: 15 b9 86 0c 00 f7 d8 64 89 02 b8 ff ff ff ff eb bf 0f 1f 44 00 00 f3 0f 1e fa 80 3d e5 08 0d 00 00 74 13 b8 2e 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 4c c3 0f 1f 00 55 48 89 e5 48 83 ec 20 89 55
 RSP: 002b:00007ffed4c35638 EFLAGS: 00000202 ORIG_RAX: 000000000000002e
 RAX: ffffffffffffffda RBX: 000055a2efcc75e0 RCX: 00007f35238bb764
 RDX: 0000000000000000 RSI: 00007ffed4c356a0 RDI: 0000000000000003
 RBP: 00007ffed4c35710 R08: 0000000000000010 R09: 00007f3523984b20
 R10: 0000000000000004 R11: 0000000000000202 R12: 00007ffed4c35790
 R13: 000000006947df8f R14: 000055a2efcc75e0 R15: 00007ffed4c35780

---
 .../net/ethernet/mellanox/mlx5/core/en_tc.c   | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

index a8773b2342c2..424786f489ec 100644
@@ -2147,11 +2147,14 @@ static void mlx5e_tc_del_fdb_peer_flow(struct mlx5e_tc_flow *flow,
 
 static void mlx5e_tc_del_fdb_peers_flow(struct mlx5e_tc_flow *flow)
 {
+	struct mlx5_devcom_comp_dev *devcom;
+	struct mlx5_devcom_comp_dev *pos;
+	struct mlx5_eswitch *peer_esw;
 	int i;
 
-	for (i = 0; i < MLX5_MAX_PORTS; i++) {
-		if (i == mlx5_get_dev_index(flow->priv->mdev))
-			continue;
+	devcom = flow->priv->mdev->priv.eswitch->devcom;
+	mlx5_devcom_for_each_peer_entry(devcom, peer_esw, pos) {
+		i = mlx5_get_dev_index(peer_esw->dev);
 		mlx5e_tc_del_fdb_peer_flow(flow, i);
 	}
 }

@@ -5513,12 +5516,16 @@ int mlx5e_tc_num_filters(struct mlx5e_priv *priv, unsigned long flags)
 
 void mlx5e_tc_clean_fdb_peer_flows(struct mlx5_eswitch *esw)
 {
+	struct mlx5_devcom_comp_dev *devcom;
+	struct mlx5_devcom_comp_dev *pos;
 	struct mlx5e_tc_flow *flow, *tmp;
+	struct mlx5_eswitch *peer_esw;
 	int i;
 
-	for (i = 0; i < MLX5_MAX_PORTS; i++) {
-		if (i == mlx5_get_dev_index(esw->dev))
-			continue;
+	devcom = esw->devcom;
+
+	mlx5_devcom_for_each_peer_entry(devcom, peer_esw, pos) {
+		i = mlx5_get_dev_index(peer_esw->dev);
 		list_for_each_entry_safe(flow, tmp, &esw->offloads.peer_flows[i], peer[i])
 			mlx5e_tc_del_fdb_peers_flow(flow);
 	}

2.40.1

From: Gal Pressman <[email protected]>

The driver's ndo_get_stats64 callback is only reporting mlx5 counters, without accounting for the netdev stats, causing errors from the network stack to be invisible in statistics.

Add netdev_stats_to_stats64() call to first populate the counters, then add mlx5 counters on top, ensuring both are accounted for (where appropriate).

---
 .../net/ethernet/mellanox/mlx5/core/en_main.c | 20 ++++++++++---------
 1 file changed, 11 insertions(+), 9 deletions(-)

index f83359f7fdea..4b2963bbe7ff 100644
@@ -4052,6 +4052,8 @@ mlx5e_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats)
 		mlx5e_queue_update_stats(priv);
 	}
 
+	netdev_stats_to_stats64(stats, &dev->stats);
+
 	if (mlx5e_is_uplink_rep(priv)) {
 		struct mlx5e_vport_stats *vstats = &priv->stats.vport;
 

@@ -4068,21 +4070,21 @@ mlx5e_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats)
 		mlx5e_fold_sw_stats64(priv, stats);
 	}
 
-	stats->rx_missed_errors = priv->stats.qcnt.rx_out_of_buffer;
-	stats->rx_dropped = PPORT_2863_GET(pstats, if_in_discards);
+	stats->rx_missed_errors += priv->stats.qcnt.rx_out_of_buffer;
+	stats->rx_dropped += PPORT_2863_GET(pstats, if_in_discards);
 
-	stats->rx_length_errors =
+	stats->rx_length_errors +=
 		PPORT_802_3_GET(pstats, a_in_range_length_errors) +
 		PPORT_802_3_GET(pstats, a_out_of_range_length_field) +
 		PPORT_802_3_GET(pstats, a_frame_too_long_errors) +
 		VNIC_ENV_GET(&priv->stats.vnic, eth_wqe_too_small);
-	stats->rx_crc_errors =
+	stats->rx_crc_errors +=
 		PPORT_802_3_GET(pstats, a_frame_check_sequence_errors);
-	stats->rx_frame_errors = PPORT_802_3_GET(pstats, a_alignment_errors);
-	stats->tx_aborted_errors = PPORT_2863_GET(pstats, if_out_discards);
-	stats->rx_errors = stats->rx_length_errors + stats->rx_crc_errors +
-			   stats->rx_frame_errors;
-	stats->tx_errors = stats->tx_aborted_errors + stats->tx_carrier_errors;
+	stats->rx_frame_errors += PPORT_802_3_GET(pstats, a_alignment_errors);
+	stats->tx_aborted_errors += PPORT_2863_GET(pstats, if_out_discards);
+	stats->rx_errors += stats->rx_length_errors + stats->rx_crc_errors +
+			    stats->rx_frame_errors;
+	stats->tx_errors += stats->tx_aborted_errors + stats->tx_carrier_errors;
 }
 
 static void mlx5e_nic_set_rx_mode(struct mlx5e_priv *priv)

2.40.1

I cooked up a patch to fix this generically in the core... but I can't actually find any "errors from the network stack" that are accounted to dev->stats. Could you be more specific about the issues you were seeing?

My original motivation was identifying packet drops in the GRE stack, specifically in gre_rcv() after an error in gre_parse_header() (in my case, due to a checksum error).
Currently, these packets are silently dropped. I have additional patches that increment the rx_dropped/rx_crc_errors counters in that path, which exposed the issue, but they haven't been submitted yet.

However, you are right that it's hard to find existing dev->stats increments, the use case this currently fixes is an error in __bpf_redirect_neigh_v4()/__bpf_redirect_neigh_v6().

Thanks,

For the series: