net: cpsw: Execute ndo_set_rx_mode callback in a work queue

These two patches resolve an RTNL assertion call trace issue in both the legacy and new cpsw drivers.

Thanks,
Kevin

---
Changes in v5:
- Apply the same fix to the legacy cpsw driver as well.

- Use goto label in cpsw_ndo_set_rx_mode_work() for the undo path as suggested by Jakub.

- Move the disable_work_sync() call to after unregister_netdev() as suggested by Jakub.

- Link to v4: https://lore.kernel.org/r/[email protected]

Changes in v4:
- Using schedule_work() instead of creating a dedicated workqueue.

- Link to v3: https://lore.kernel.org/r/[email protected]

Changes in v3:
- Resolve the deadlock issue identified in the AI review [2] by moving the netif_running() check under the RTNL lock and removing the cancel_work_sync() call in cpsw_ndo_stop().

- Link to v2: https://lore.kernel.org/r/[email protected]

Changes in v2:
- Addresses the issue identified in the AI review [1]:
- Adds a netif_running() check in cpsw_ndo_set_rx_mode_work()
- Cancels the rx_mode_work in cpsw_ndo_stop()

- Link to v1: https://lore.kernel.org/r/[email protected]

[1] https://netdev-ai.bots.linux.dev/ai-review.html?id=bd885e1e-1aed-4755-ad60-7150737ad0f5
[2] https://netdev-ai.bots.linux.dev/ai-review.html?id=c9fc3cf8-a06c-4cb8-b26b-910e775951a0

---
Kevin Hao (2):
net: cpsw_new: Execute ndo_set_rx_mode callback in a work queue
net: cpsw: Execute ndo_set_rx_mode callback in a work queue

 drivers/net/ethernet/ti/cpsw.c      | 41 +++++++++++++++++++++++++++++++------
 drivers/net/ethernet/ti/cpsw_new.c  | 34 +++++++++++++++++++++++++-----
 drivers/net/ethernet/ti/cpsw_priv.h |  1 +
 3 files changed, 65 insertions(+), 11 deletions(-)

---
base-commit: 193579fe01389bc21aff0051d13f24e8ea95b47d
change-id: 20260123-bbb-dc3675f671d0

Best regards,

Kevin Hao <[email protected]>

Commit 1767bb2d47b7 ("ipv6: mcast: Don't hold RTNL for IPV6_ADD_MEMBERSHIP and MCAST_JOIN_GROUP.") removed the RTNL lock for IPV6_ADD_MEMBERSHIP and MCAST_JOIN_GROUP operations. However, this change triggered the following call trace on my BeagleBone Black board:

  WARNING: net/8021q/vlan_core.c:236 at vlan_for_each+0x120/0x124, CPU#0: rpcbind/496
  RTNL: assertion failed at net/8021q/vlan_core.c (236)
  Modules linked in:
  CPU: 0 UID: 997 PID: 496 Comm: rpcbind Not tainted 6.19.0-rc6-next-20260122-yocto-standard+ #8 PREEMPT
  Hardware name: Generic AM33XX (Flattened Device Tree)
  Call trace:
   unwind_backtrace from show_stack+0x28/0x2c
   show_stack from dump_stack_lvl+0x30/0x38
   dump_stack_lvl from __warn+0xb8/0x11c
   __warn from warn_slowpath_fmt+0x130/0x194
   warn_slowpath_fmt from vlan_for_each+0x120/0x124
   vlan_for_each from cpsw_add_mc_addr+0x54/0xd8
   cpsw_add_mc_addr from __hw_addr_ref_sync_dev+0xc4/0xec
   __hw_addr_ref_sync_dev from __dev_mc_add+0x78/0x88
   __dev_mc_add from igmp6_group_added+0x84/0xec
   igmp6_group_added from __ipv6_dev_mc_inc+0x1fc/0x2f0
   __ipv6_dev_mc_inc from __ipv6_sock_mc_join+0x124/0x1b4
   __ipv6_sock_mc_join from do_ipv6_setsockopt+0x84c/0x1168
   do_ipv6_setsockopt from ipv6_setsockopt+0x88/0xc8
   ipv6_setsockopt from do_sock_setsockopt+0xe8/0x19c
   do_sock_setsockopt from __sys_setsockopt+0x84/0xac
   __sys_setsockopt from ret_fast_syscall+0x0/0x5

This trace occurs because vlan_for_each() is called within cpsw_ndo_set_rx_mode(), which expects the RTNL lock to be held. Since modifying vlan_for_each() to operate without the RTNL lock is not straightforward, and because ndo_set_rx_mode() is invoked both with and without the RTNL lock across different code paths, simply adding rtnl_lock() in cpsw_ndo_set_rx_mode() is not a viable solution.

To resolve this issue, we opt to execute the actual processing within a work queue, following the approach used by the icssg-prueth driver.

---
Cc: Siddharth Vadapalli <[email protected]>
Cc: Roger Quadros <[email protected]>
Cc: Andrew Lunn <[email protected]>
Cc: "David S. Miller" <[email protected]>
Cc: Eric Dumazet <[email protected]>
Cc: Jakub Kicinski <[email protected]>
Cc: Paolo Abeni <[email protected]>
Cc: Vladimir Oltean <[email protected]>
Cc: Kuniyuki Iwashima <[email protected]>
Cc: [email protected]

---
 drivers/net/ethernet/ti/cpsw_new.c  | 34 +++++++++++++++++++++++++++++-----
 drivers/net/ethernet/ti/cpsw_priv.h |  1 +
 2 files changed, 30 insertions(+), 5 deletions(-)

index ab88d4c02cbde76207f89cf433e2b383dcde6a83..21af0a10626aaf0ce6ecb04837899801743f3894 100644
@@ -248,16 +248,22 @@ static int cpsw_purge_all_mc(struct net_device *ndev, const u8 *addr, int num)
 	return 0;
 }
 
-static void cpsw_ndo_set_rx_mode(struct net_device *ndev)
+static void cpsw_ndo_set_rx_mode_work(struct work_struct *work)
 {
-	struct cpsw_priv *priv = netdev_priv(ndev);
+	struct cpsw_priv *priv = container_of(work, struct cpsw_priv, rx_mode_work);
 	struct cpsw_common *cpsw = priv->cpsw;
+	struct net_device *ndev = priv->ndev;
 
+	rtnl_lock();
+	if (!netif_running(ndev))
+		goto unlock_rtnl;
+
+	netif_addr_lock_bh(ndev);
 	if (ndev->flags & IFF_PROMISC) {
 		/* Enable promiscuous mode */
 		cpsw_set_promiscious(ndev, true);
 		cpsw_ale_set_allmulti(cpsw->ale, IFF_ALLMULTI, priv->emac_port);
-		return;
+		goto unlock_addr;
 	}
 
 	/* Disable promiscuous mode */

@@ -270,6 +276,18 @@ static void cpsw_ndo_set_rx_mode(struct net_device *ndev)
 	/* add/remove mcast address either for real netdev or for vlan */
 	__hw_addr_ref_sync_dev(&ndev->mc, ndev, cpsw_add_mc_addr,
 			       cpsw_del_mc_addr);
+
+unlock_addr:
+	netif_addr_unlock_bh(ndev);
+unlock_rtnl:
+	rtnl_unlock();
+}
+
+static void cpsw_ndo_set_rx_mode(struct net_device *ndev)
+{
+	struct cpsw_priv *priv = netdev_priv(ndev);
+
+	schedule_work(&priv->rx_mode_work);
 }
 
 static unsigned int cpsw_rxbuf_total_len(unsigned int len)

@@ -1398,6 +1416,7 @@ static int cpsw_create_ports(struct cpsw_common *cpsw)
 		priv->msg_enable = netif_msg_init(debug_level, CPSW_DEBUG);
 		priv->emac_port = i + 1;
 		priv->tx_packet_min = CPSW_MIN_PACKET_SIZE;
+		INIT_WORK(&priv->rx_mode_work, cpsw_ndo_set_rx_mode_work);
 
 		if (is_valid_ether_addr(slave_data->mac_addr)) {
 			ether_addr_copy(priv->mac_addr, slave_data->mac_addr);

@@ -1447,13 +1466,18 @@ static int cpsw_create_ports(struct cpsw_common *cpsw)
 
 static void cpsw_unregister_ports(struct cpsw_common *cpsw)
 {
+	struct net_device *ndev;
+	struct cpsw_priv *priv;
 	int i = 0;
 
 	for (i = 0; i < cpsw->data.slaves; i++) {
-		if (!cpsw->slaves[i].ndev)
+		ndev = cpsw->slaves[i].ndev;
+		if (!ndev)
 			continue;
 
-		unregister_netdev(cpsw->slaves[i].ndev);
+		priv = netdev_priv(ndev);
+		unregister_netdev(ndev);
+		disable_work_sync(&priv->rx_mode_work);
 	}
 }
 

index 91add8925e235c6cf5542fde11f3383b9234c872..acb6181c5c9e1bf6ed46a7fd14ce422efc0b724e 100644
@@ -391,6 +391,7 @@ struct cpsw_priv {
 	u32 tx_packet_min;
 	struct cpsw_ale_ratelimit ale_bc_ratelimit;
 	struct cpsw_ale_ratelimit ale_mc_ratelimit;
+	struct work_struct rx_mode_work;
 };
 
 #define ndev_to_cpsw(ndev) (((struct cpsw_priv *)netdev_priv(ndev))->cpsw)

2.52.0

Commit 1767bb2d47b7 ("ipv6: mcast: Don't hold RTNL for IPV6_ADD_MEMBERSHIP and MCAST_JOIN_GROUP.") removed the RTNL lock for IPV6_ADD_MEMBERSHIP and MCAST_JOIN_GROUP operations. However, this change triggered the following call trace on my BeagleBone Black board:

  WARNING: net/8021q/vlan_core.c:236 at vlan_for_each+0x120/0x124, CPU#0: rpcbind/481
  RTNL: assertion failed at net/8021q/vlan_core.c (236)
  Modules linked in:
  CPU: 0 UID: 997 PID: 481 Comm: rpcbind Not tainted 6.19.0-rc7-next-20260130-yocto-standard+ #35 PREEMPT
  Hardware name: Generic AM33XX (Flattened Device Tree)
  Call trace:
   unwind_backtrace from show_stack+0x28/0x2c
   show_stack from dump_stack_lvl+0x30/0x38
   dump_stack_lvl from __warn+0xb8/0x11c
   __warn from warn_slowpath_fmt+0x130/0x194
   warn_slowpath_fmt from vlan_for_each+0x120/0x124
   vlan_for_each from cpsw_add_mc_addr+0x54/0x98
   cpsw_add_mc_addr from __hw_addr_ref_sync_dev+0xc4/0xec
   __hw_addr_ref_sync_dev from __dev_mc_add+0x78/0x88
   __dev_mc_add from igmp6_group_added+0x84/0xec
   igmp6_group_added from __ipv6_dev_mc_inc+0x1fc/0x2f0
   __ipv6_dev_mc_inc from __ipv6_sock_mc_join+0x124/0x1b4
   __ipv6_sock_mc_join from do_ipv6_setsockopt+0x84c/0x1168
   do_ipv6_setsockopt from ipv6_setsockopt+0x88/0xc8
   ipv6_setsockopt from do_sock_setsockopt+0xe8/0x19c
   do_sock_setsockopt from __sys_setsockopt+0x84/0xac
   __sys_setsockopt from ret_fast_syscall+0x0/0x54

This trace occurs because vlan_for_each() is called within cpsw_ndo_set_rx_mode(), which expects the RTNL lock to be held. Since modifying vlan_for_each() to operate without the RTNL lock is not straightforward, and because ndo_set_rx_mode() is invoked both with and without the RTNL lock across different code paths, simply adding rtnl_lock() in cpsw_ndo_set_rx_mode() is not a viable solution.

To resolve this issue, we opt to execute the actual processing within a work queue, following the approach used by the icssg-prueth driver.

Please note: To reproduce this issue, I manually reverted the changes to am335x-bone-common.dtsi from commit c477358e66a3 ("ARM: dts: am335x-bone:
switch to new cpsw switch drv") in order to revert to the legacy cpsw driver.

---
Cc: Siddharth Vadapalli <[email protected]>
Cc: Roger Quadros <[email protected]>
Cc: Andrew Lunn <[email protected]>
Cc: "David S. Miller" <[email protected]>
Cc: Eric Dumazet <[email protected]>
Cc: Jakub Kicinski <[email protected]>
Cc: Paolo Abeni <[email protected]>
Cc: Vladimir Oltean <[email protected]>
Cc: Kuniyuki Iwashima <[email protected]>
Cc: [email protected]

---
 drivers/net/ethernet/ti/cpsw.c | 41 +++++++++++++++++++++++++++++++++++------
 1 file changed, 35 insertions(+), 6 deletions(-)

index 54c24cd3d3be639c90111856efaf154cd77f7ee4..b0e18bdc2c851086d6db6a3d7b9659c23c73b969 100644
@@ -305,12 +305,19 @@ static int cpsw_purge_all_mc(struct net_device *ndev, const u8 *addr, int num)
 	return 0;
 }
 
-static void cpsw_ndo_set_rx_mode(struct net_device *ndev)
+static void cpsw_ndo_set_rx_mode_work(struct work_struct *work)
 {
-	struct cpsw_priv *priv = netdev_priv(ndev);
+	struct cpsw_priv *priv = container_of(work, struct cpsw_priv, rx_mode_work);
 	struct cpsw_common *cpsw = priv->cpsw;
+	struct net_device *ndev = priv->ndev;
 	int slave_port = -1;
 
+	rtnl_lock();
+	if (!netif_running(ndev))
+		goto unlock_rtnl;
+
+	netif_addr_lock_bh(ndev);
+
 	if (cpsw->data.dual_emac)
 		slave_port = priv->emac_port + 1;
 

@@ -318,7 +325,7 @@ static void cpsw_ndo_set_rx_mode(struct net_device *ndev)
 		/* Enable promiscuous mode */
 		cpsw_set_promiscious(ndev, true);
 		cpsw_ale_set_allmulti(cpsw->ale, IFF_ALLMULTI, slave_port);
-		return;
+		goto unlock_addr;
 	} else {
 		/* Disable promiscuous mode */
 		cpsw_set_promiscious(ndev, false);

@@ -331,6 +338,18 @@ static void cpsw_ndo_set_rx_mode(struct net_device *ndev)
 	/* add/remove mcast address either for real netdev or for vlan */
 	__hw_addr_ref_sync_dev(&ndev->mc, ndev, cpsw_add_mc_addr,
 			       cpsw_del_mc_addr);
+
+unlock_addr:
+	netif_addr_unlock_bh(ndev);
+unlock_rtnl:
+	rtnl_unlock();
+}
+
+static void cpsw_ndo_set_rx_mode(struct net_device *ndev)
+{
+	struct cpsw_priv *priv = netdev_priv(ndev);
+
+	schedule_work(&priv->rx_mode_work);
 }
 
 static unsigned int cpsw_rxbuf_total_len(unsigned int len)

@@ -1472,6 +1491,7 @@ static int cpsw_probe_dual_emac(struct cpsw_priv *priv)
 	priv_sl2->ndev = ndev;
 	priv_sl2->dev  = &ndev->dev;
 	priv_sl2->msg_enable = netif_msg_init(debug_level, CPSW_DEBUG);
+	INIT_WORK(&priv_sl2->rx_mode_work, cpsw_ndo_set_rx_mode_work);
 
 	if (is_valid_ether_addr(data->slave_data[1].mac_addr)) {
 		memcpy(priv_sl2->mac_addr, data->slave_data[1].mac_addr,

@@ -1653,6 +1673,7 @@ static int cpsw_probe(struct platform_device *pdev)
 	priv->dev  = dev;
 	priv->msg_enable = netif_msg_init(debug_level, CPSW_DEBUG);
 	priv->emac_port = 0;
+	INIT_WORK(&priv->rx_mode_work, cpsw_ndo_set_rx_mode_work);
 
 	if (is_valid_ether_addr(data->slave_data[0].mac_addr)) {
 		memcpy(priv->mac_addr, data->slave_data[0].mac_addr, ETH_ALEN);

@@ -1758,6 +1779,8 @@ static int cpsw_probe(struct platform_device *pdev)
 static void cpsw_remove(struct platform_device *pdev)
 {
 	struct cpsw_common *cpsw = platform_get_drvdata(pdev);
+	struct net_device *ndev;
+	struct cpsw_priv *priv;
 	int i, ret;
 
 	ret = pm_runtime_resume_and_get(&pdev->dev);

@@ -1770,9 +1793,15 @@ static void cpsw_remove(struct platform_device *pdev)
 		return;
 	}
 
-	for (i = 0; i < cpsw->data.slaves; i++)
-		if (cpsw->slaves[i].ndev)
-			unregister_netdev(cpsw->slaves[i].ndev);
+	for (i = 0; i < cpsw->data.slaves; i++) {
+		ndev = cpsw->slaves[i].ndev;
+		if (!ndev)
+			continue;
+
+		priv = netdev_priv(ndev);
+		unregister_netdev(ndev);
+		disable_work_sync(&priv->rx_mode_work);
+	}
 
 	cpts_release(cpsw->cpts);
 	cpdma_ctlr_destroy(cpsw->dma);

2.52.0