[PATCH net 0/2] net: cpsw_new: Fix multiple issues in the cpsw_probe() error path

[PATCH net 0/2] net: cpsw_new: Fix multiple issues in the cpsw_probe() error path

[Kevin Hao]

These two patches address duplicate or unnecessary netdev unregistration
in the cpsw_probe() error handling path.

---
Cc: Siddharth Vadapalli <s-vadapalli@ti.com>
Cc: Roger Quadros <rogerq@kernel.org>
Cc: Andrew Lunn <andrew+netdev@lunn.ch>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Vladimir Oltean <vladimir.oltean@nxp.com>
Cc: Saeed Mahameed <saeedm@nvidia.com>
Cc: Daniel Zahka <daniel.zahka@gmail.com>
Cc: Lorenzo Bianconi <lorenzo@kernel.org>
Cc: Alexander Sverdlin <alexander.sverdlin@gmail.com>
Cc: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Cc: Murali Karicheri <m-karicheri2@ti.com>
Cc: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Cc: Grygorii Strashko <grygorii.strashko@ti.com>
Cc: linux-omap@vger.kernel.org
Cc: stable@vger.kernel.org

---
Kevin Hao (2):
      net: cpsw_new: Fix unnecessary netdev unregistration in cpsw_probe() error path
      net: cpsw_new: Fix potential unregister of netdev that has not been registered yet

 drivers/net/ethernet/ti/cpsw_new.c | 15 ++++++---------
 1 file changed, 6 insertions(+), 9 deletions(-)
---
base-commit: 0f8a890c4524d6e4013ff225e70de2aed7e6d726
change-id: 20260202-cpsw-error-path-dd5bc72d6daa

Best regards,
-- 
Kevin Hao <haokexin@gmail.com>

[PATCH net 1/2] net: cpsw_new: Fix unnecessary netdev unregistration in cpsw_probe() error path

[Kevin Hao]

The current error handling in cpsw_probe() has two issues:
- cpsw_unregister_ports() may be called before cpsw_register_ports() has
  been executed.

- cpsw_unregister_ports() is already invoked within cpsw_register_ports()
  in case of a register_netdev() failure, but the error path would call
  it again.

Fixes: ed3525eda4c4 ("net: ethernet: ti: introduce cpsw switchdev based driver part 1 - dual-emac")
Signed-off-by: Kevin Hao <haokexin@gmail.com>
Cc: stable@vger.kernel.org
---
 drivers/net/ethernet/ti/cpsw_new.c | 12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/ti/cpsw_new.c b/drivers/net/ethernet/ti/cpsw_new.c
index 21af0a10626aaf0ce6ecb04837899801743f3894..b9fc31eb06134dae33427eaba06341c39eb4b41c 100644
--- a/drivers/net/ethernet/ti/cpsw_new.c
+++ b/drivers/net/ethernet/ti/cpsw_new.c
@@ -2003,7 +2003,7 @@ static int cpsw_probe(struct platform_device *pdev)
 	/* setup netdevs */
 	ret = cpsw_create_ports(cpsw);
 	if (ret)
-		goto clean_unregister_netdev;
+		goto clean_cpts;
 
 	/* Grab RX and TX IRQs. Note that we also have RX_THRESHOLD and
 	 * MISC IRQs which are always kept disabled with this driver so
@@ -2017,14 +2017,14 @@ static int cpsw_probe(struct platform_device *pdev)
 			       0, dev_name(dev), cpsw);
 	if (ret < 0) {
 		dev_err(dev, "error attaching irq (%d)\n", ret);
-		goto clean_unregister_netdev;
+		goto clean_cpts;
 	}
 
 	ret = devm_request_irq(dev, cpsw->irqs_table[1], cpsw_tx_interrupt,
 			       0, dev_name(dev), cpsw);
 	if (ret < 0) {
 		dev_err(dev, "error attaching irq (%d)\n", ret);
-		goto clean_unregister_netdev;
+		goto clean_cpts;
 	}
 
 	if (!cpsw->cpts)
@@ -2034,7 +2034,7 @@ static int cpsw_probe(struct platform_device *pdev)
 			       0, dev_name(&pdev->dev), cpsw);
 	if (ret < 0) {
 		dev_err(dev, "error attaching misc irq (%d)\n", ret);
-		goto clean_unregister_netdev;
+		goto clean_cpts;
 	}
 
 	/* Enable misc CPTS evnt_pend IRQ */
@@ -2043,7 +2043,7 @@ static int cpsw_probe(struct platform_device *pdev)
 skip_cpts:
 	ret = cpsw_register_notifiers(cpsw);
 	if (ret)
-		goto clean_unregister_netdev;
+		goto clean_cpts;
 
 	ret = cpsw_register_devlink(cpsw);
 	if (ret)
@@ -2065,8 +2065,6 @@ static int cpsw_probe(struct platform_device *pdev)
 
 clean_unregister_notifiers:
 	cpsw_unregister_notifiers(cpsw);
-clean_unregister_netdev:
-	cpsw_unregister_ports(cpsw);
 clean_cpts:
 	cpts_release(cpsw->cpts);
 	cpdma_ctlr_destroy(cpsw->dma);

-- 
2.52.0

Re: [PATCH net 1/2] net: cpsw_new: Fix unnecessary netdev unregistration in cpsw_probe() error path

[Alexander Sverdlin]

On Thu, 2026-02-05 at 10:47 +0800, Kevin Hao wrote:
> The current error handling in cpsw_probe() has two issues:
> - cpsw_unregister_ports() may be called before cpsw_register_ports() has
>   been executed.
> 
> - cpsw_unregister_ports() is already invoked within cpsw_register_ports()
>   in case of a register_netdev() failure, but the error path would call
>   it again.
> 
> Fixes: ed3525eda4c4 ("net: ethernet: ti: introduce cpsw switchdev based driver part 1 - dual-emac")
> Signed-off-by: Kevin Hao <haokexin@gmail.com>

Reviewed-by: Alexander Sverdlin <alexander.sverdlin@gmail.com>

> Cc: stable@vger.kernel.org
> ---
>  drivers/net/ethernet/ti/cpsw_new.c | 12 +++++-------
>  1 file changed, 5 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/net/ethernet/ti/cpsw_new.c b/drivers/net/ethernet/ti/cpsw_new.c
> index 21af0a10626aaf0ce6ecb04837899801743f3894..b9fc31eb06134dae33427eaba06341c39eb4b41c 100644
> --- a/drivers/net/ethernet/ti/cpsw_new.c
> +++ b/drivers/net/ethernet/ti/cpsw_new.c
> @@ -2003,7 +2003,7 @@ static int cpsw_probe(struct platform_device *pdev)
>  	/* setup netdevs */
>  	ret = cpsw_create_ports(cpsw);
>  	if (ret)
> -		goto clean_unregister_netdev;
> +		goto clean_cpts;
>  
>  	/* Grab RX and TX IRQs. Note that we also have RX_THRESHOLD and
>  	 * MISC IRQs which are always kept disabled with this driver so
> @@ -2017,14 +2017,14 @@ static int cpsw_probe(struct platform_device *pdev)
>  			       0, dev_name(dev), cpsw);
>  	if (ret < 0) {
>  		dev_err(dev, "error attaching irq (%d)\n", ret);
> -		goto clean_unregister_netdev;
> +		goto clean_cpts;
>  	}
>  
>  	ret = devm_request_irq(dev, cpsw->irqs_table[1], cpsw_tx_interrupt,
>  			       0, dev_name(dev), cpsw);
>  	if (ret < 0) {
>  		dev_err(dev, "error attaching irq (%d)\n", ret);
> -		goto clean_unregister_netdev;
> +		goto clean_cpts;
>  	}
>  
>  	if (!cpsw->cpts)
> @@ -2034,7 +2034,7 @@ static int cpsw_probe(struct platform_device *pdev)
>  			       0, dev_name(&pdev->dev), cpsw);
>  	if (ret < 0) {
>  		dev_err(dev, "error attaching misc irq (%d)\n", ret);
> -		goto clean_unregister_netdev;
> +		goto clean_cpts;
>  	}
>  
>  	/* Enable misc CPTS evnt_pend IRQ */
> @@ -2043,7 +2043,7 @@ static int cpsw_probe(struct platform_device *pdev)
>  skip_cpts:
>  	ret = cpsw_register_notifiers(cpsw);
>  	if (ret)
> -		goto clean_unregister_netdev;
> +		goto clean_cpts;
>  
>  	ret = cpsw_register_devlink(cpsw);
>  	if (ret)
> @@ -2065,8 +2065,6 @@ static int cpsw_probe(struct platform_device *pdev)
>  
>  clean_unregister_notifiers:
>  	cpsw_unregister_notifiers(cpsw);
> -clean_unregister_netdev:
> -	cpsw_unregister_ports(cpsw);
>  clean_cpts:
>  	cpts_release(cpsw->cpts);
>  	cpdma_ctlr_destroy(cpsw->dma);

-- 
Alexander Sverdlin.

[PATCH net 2/2] net: cpsw_new: Fix potential unregister of netdev that has not been registered yet

[Kevin Hao]

If an error occurs during register_netdev() for the first MAC in
cpsw_register_ports(), even though cpsw->slaves[0].ndev is set to NULL,
cpsw->slaves[1].ndev would remain unchanged. This could later cause
cpsw_unregister_ports() to attempt unregistering the second MAC.
To address this, add a check for ndev->reg_state before calling
unregister_netdev(). With this change, setting cpsw->slaves[i].ndev
to NULL becomes unnecessary and can be removed accordingly.

Fixes: ed3525eda4c4 ("net: ethernet: ti: introduce cpsw switchdev based driver part 1 - dual-emac")
Signed-off-by: Kevin Hao <haokexin@gmail.com>
Cc: stable@vger.kernel.org
---
 drivers/net/ethernet/ti/cpsw_new.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/ti/cpsw_new.c b/drivers/net/ethernet/ti/cpsw_new.c
index b9fc31eb06134dae33427eaba06341c39eb4b41c..7f42f58a4b031fab4c93680c153383e8eeb8f7f8 100644
--- a/drivers/net/ethernet/ti/cpsw_new.c
+++ b/drivers/net/ethernet/ti/cpsw_new.c
@@ -1472,7 +1472,7 @@ static void cpsw_unregister_ports(struct cpsw_common *cpsw)
 
 	for (i = 0; i < cpsw->data.slaves; i++) {
 		ndev = cpsw->slaves[i].ndev;
-		if (!ndev)
+		if (!ndev || ndev->reg_state != NETREG_REGISTERED)
 			continue;
 
 		priv = netdev_priv(ndev);
@@ -1494,7 +1494,6 @@ static int cpsw_register_ports(struct cpsw_common *cpsw)
 		if (ret) {
 			dev_err(cpsw->dev,
 				"cpsw: err registering net device%d\n", i);
-			cpsw->slaves[i].ndev = NULL;
 			break;
 		}
 	}

-- 
2.52.0

Re: [PATCH net 2/2] net: cpsw_new: Fix potential unregister of netdev that has not been registered yet

[Alexander Sverdlin]

On Thu, 2026-02-05 at 10:47 +0800, Kevin Hao wrote:
> If an error occurs during register_netdev() for the first MAC in
> cpsw_register_ports(), even though cpsw->slaves[0].ndev is set to NULL,
> cpsw->slaves[1].ndev would remain unchanged. This could later cause
> cpsw_unregister_ports() to attempt unregistering the second MAC.
> To address this, add a check for ndev->reg_state before calling
> unregister_netdev(). With this change, setting cpsw->slaves[i].ndev
> to NULL becomes unnecessary and can be removed accordingly.
> 
> Fixes: ed3525eda4c4 ("net: ethernet: ti: introduce cpsw switchdev based driver part 1 - dual-emac")
> Signed-off-by: Kevin Hao <haokexin@gmail.com>

Reviewed-by: Alexander Sverdlin <alexander.sverdlin@gmail.com>

> Cc: stable@vger.kernel.org
> ---
>  drivers/net/ethernet/ti/cpsw_new.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/drivers/net/ethernet/ti/cpsw_new.c b/drivers/net/ethernet/ti/cpsw_new.c
> index b9fc31eb06134dae33427eaba06341c39eb4b41c..7f42f58a4b031fab4c93680c153383e8eeb8f7f8 100644
> --- a/drivers/net/ethernet/ti/cpsw_new.c
> +++ b/drivers/net/ethernet/ti/cpsw_new.c
> @@ -1472,7 +1472,7 @@ static void cpsw_unregister_ports(struct cpsw_common *cpsw)
>  
>  	for (i = 0; i < cpsw->data.slaves; i++) {
>  		ndev = cpsw->slaves[i].ndev;
> -		if (!ndev)
> +		if (!ndev || ndev->reg_state != NETREG_REGISTERED)
>  			continue;
>  
>  		priv = netdev_priv(ndev);
> @@ -1494,7 +1494,6 @@ static int cpsw_register_ports(struct cpsw_common *cpsw)
>  		if (ret) {
>  			dev_err(cpsw->dev,
>  				"cpsw: err registering net device%d\n", i);
> -			cpsw->slaves[i].ndev = NULL;
>  			break;
>  		}
>  	}

-- 
Alexander Sverdlin.

Re: [PATCH net 0/2] net: cpsw_new: Fix multiple issues in the cpsw_probe() error path

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Thu, 05 Feb 2026 10:47:01 +0800 you wrote:
> These two patches address duplicate or unnecessary netdev unregistration
> in the cpsw_probe() error handling path.
> 
> ---
> Cc: Siddharth Vadapalli <s-vadapalli@ti.com>
> Cc: Roger Quadros <rogerq@kernel.org>
> Cc: Andrew Lunn <andrew+netdev@lunn.ch>
> Cc: "David S. Miller" <davem@davemloft.net>
> Cc: Eric Dumazet <edumazet@google.com>
> Cc: Jakub Kicinski <kuba@kernel.org>
> Cc: Paolo Abeni <pabeni@redhat.com>
> Cc: Vladimir Oltean <vladimir.oltean@nxp.com>
> Cc: Saeed Mahameed <saeedm@nvidia.com>
> Cc: Daniel Zahka <daniel.zahka@gmail.com>
> Cc: Lorenzo Bianconi <lorenzo@kernel.org>
> Cc: Alexander Sverdlin <alexander.sverdlin@gmail.com>
> Cc: Nicolas Dichtel <nicolas.dichtel@6wind.com>
> Cc: Murali Karicheri <m-karicheri2@ti.com>
> Cc: Ilias Apalodimas <ilias.apalodimas@linaro.org>
> Cc: Grygorii Strashko <grygorii.strashko@ti.com>
> Cc: linux-omap@vger.kernel.org
> Cc: stable@vger.kernel.org
> 
> [...]

Here is the summary with links:
  - [net,1/2] net: cpsw_new: Fix unnecessary netdev unregistration in cpsw_probe() error path
    https://git.kernel.org/netdev/net/c/62db84b7efa6
  - [net,2/2] net: cpsw_new: Fix potential unregister of netdev that has not been registered yet
    https://git.kernel.org/netdev/net/c/9d724b34fbe1

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html