[PATCH net v2 1/3] bus: fsl-mc: Fix potential double device reference in fsl_mc_get_endpoint()

[PATCH net v2 1/3] bus: fsl-mc: Fix potential double device reference in fsl_mc_get_endpoint()

[Ma Ke]

The fsl_mc_get_endpoint() function may call fsl_mc_device_lookup() 
twice, which would increment the device's reference count twice if 
both lookups find a device. This could lead to a reference count leak.

Found by code review.

Cc: stable@vger.kernel.org
Fixes: 1ac210d128ef ("bus: fsl-mc: add the fsl_mc_get_endpoint function")
Signed-off-by: Ma Ke <make24@iscas.ac.cn>
---
 drivers/bus/fsl-mc/fsl-mc-bus.c | 19 +++++++++----------
 1 file changed, 9 insertions(+), 10 deletions(-)

diff --git a/drivers/bus/fsl-mc/fsl-mc-bus.c b/drivers/bus/fsl-mc/fsl-mc-bus.c
index 7671bd158545..c1c0a4759c7e 100644
--- a/drivers/bus/fsl-mc/fsl-mc-bus.c
+++ b/drivers/bus/fsl-mc/fsl-mc-bus.c
@@ -943,6 +943,7 @@ struct fsl_mc_device *fsl_mc_get_endpoint(struct fsl_mc_device *mc_dev,
 	struct fsl_mc_obj_desc endpoint_desc = {{ 0 }};
 	struct dprc_endpoint endpoint1 = {{ 0 }};
 	struct dprc_endpoint endpoint2 = {{ 0 }};
+	struct fsl_mc_bus *mc_bus;
 	int state, err;
 
 	mc_bus_dev = to_fsl_mc_device(mc_dev->dev.parent);
@@ -966,6 +967,8 @@ struct fsl_mc_device *fsl_mc_get_endpoint(struct fsl_mc_device *mc_dev,
 	strcpy(endpoint_desc.type, endpoint2.type);
 	endpoint_desc.id = endpoint2.id;
 	endpoint = fsl_mc_device_lookup(&endpoint_desc, mc_bus_dev);
+	if (endpoint)
+		return endpoint;
 
 	/*
 	 * We know that the device has an endpoint because we verified by
@@ -973,17 +976,13 @@ struct fsl_mc_device *fsl_mc_get_endpoint(struct fsl_mc_device *mc_dev,
 	 * yet discovered by the fsl-mc bus, thus the lookup returned NULL.
 	 * Force a rescan of the devices in this container and retry the lookup.
 	 */
-	if (!endpoint) {
-		struct fsl_mc_bus *mc_bus = to_fsl_mc_bus(mc_bus_dev);
-
-		if (mutex_trylock(&mc_bus->scan_mutex)) {
-			err = dprc_scan_objects(mc_bus_dev, true);
-			mutex_unlock(&mc_bus->scan_mutex);
-		}
-
-		if (err < 0)
-			return ERR_PTR(err);
+	mc_bus = to_fsl_mc_bus(mc_bus_dev);
+	if (mutex_trylock(&mc_bus->scan_mutex)) {
+		err = dprc_scan_objects(mc_bus_dev, true);
+		mutex_unlock(&mc_bus->scan_mutex);
 	}
+	if (err < 0)
+		return ERR_PTR(err);
 
 	endpoint = fsl_mc_device_lookup(&endpoint_desc, mc_bus_dev);
 	/*
-- 
2.25.1

[PATCH net v2 2/3] dpaa2-eth: Fix device reference count leak in MAC endpoint handling

[Ma Ke]

The fsl_mc_get_endpoint() function uses device_find_child() for
localization, which implicitly calls get_device() to increment the
device's reference count before returning the pointer. However, the
caller dpaa2_eth_connect_mac() fails to properly release this 
reference in multiple scenarios. We should call put_device() to 
decrement reference count properly.

As comment of device_find_child() says, 'NOTE: you will need to drop
the reference with put_device() after use'.

Found by code review.

Cc: stable@vger.kernel.org
Fixes: 719479230893 ("dpaa2-eth: add MAC/PHY support through phylink")
Signed-off-by: Ma Ke <make24@iscas.ac.cn>
---
 drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
index b82f121cadad..0f4efd505332 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
@@ -4666,12 +4666,19 @@ static int dpaa2_eth_connect_mac(struct dpaa2_eth_priv *priv)
 		return PTR_ERR(dpmac_dev);
 	}
 
-	if (IS_ERR(dpmac_dev) || dpmac_dev->dev.type != &fsl_mc_bus_dpmac_type)
+	if (IS_ERR(dpmac_dev))
 		return 0;
 
+	if (dpmac_dev->dev.type != &fsl_mc_bus_dpmac_type) {
+		err = 0;
+		goto out_put_device;
+	}
+
 	mac = kzalloc(sizeof(struct dpaa2_mac), GFP_KERNEL);
-	if (!mac)
-		return -ENOMEM;
+	if (!mac) {
+		err = -ENOMEM;
+		goto out_put_device;
+	}
 
 	mac->mc_dev = dpmac_dev;
 	mac->mc_io = priv->mc_io;
@@ -4705,6 +4712,8 @@ static int dpaa2_eth_connect_mac(struct dpaa2_eth_priv *priv)
 	dpaa2_mac_close(mac);
 err_free_mac:
 	kfree(mac);
+out_put_device:
+	put_device(&dpmac_dev->dev);
 	return err;
 }
 
-- 
2.25.1

[PATCH net v2 3/3] dpaa2-switch: Fix device reference count leak in MAC endpoint handling

[Ma Ke]

The fsl_mc_get_endpoint() function uses device_find_child() for
localization, which implicitly calls get_device() to increment the
device's reference count before returning the pointer. However, the
caller dpaa2_switch_port_connect_mac() fails to properly release this 
reference in multiple scenarios. We should call put_device() to 
decrement reference count properly.

As comment of device_find_child() says, 'NOTE: you will need to drop
the reference with put_device() after use'.

Found by code review.

Cc: stable@vger.kernel.org
Fixes: 84cba72956fd ("dpaa2-switch: integrate the MAC endpoint support")
Signed-off-by: Ma Ke <make24@iscas.ac.cn>
---
 .../net/ethernet/freescale/dpaa2/dpaa2-switch.c   | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-switch.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-switch.c
index 147a93bf9fa9..4643a3380618 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-switch.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-switch.c
@@ -1448,12 +1448,19 @@ static int dpaa2_switch_port_connect_mac(struct ethsw_port_priv *port_priv)
 	if (PTR_ERR(dpmac_dev) == -EPROBE_DEFER)
 		return PTR_ERR(dpmac_dev);
 
-	if (IS_ERR(dpmac_dev) || dpmac_dev->dev.type != &fsl_mc_bus_dpmac_type)
+	if (IS_ERR(dpmac_dev))
 		return 0;
 
+	if (dpmac_dev->dev.type != &fsl_mc_bus_dpmac_type) {
+		err = 0;
+		goto out_put_device;
+	}
+
 	mac = kzalloc(sizeof(*mac), GFP_KERNEL);
-	if (!mac)
-		return -ENOMEM;
+	if (!mac) {
+		err = -ENOMEM;
+		goto out_put_device;
+	}
 
 	mac->mc_dev = dpmac_dev;
 	mac->mc_io = port_priv->ethsw_data->mc_io;
@@ -1483,6 +1490,8 @@ static int dpaa2_switch_port_connect_mac(struct ethsw_port_priv *port_priv)
 	dpaa2_mac_close(mac);
 err_free_mac:
 	kfree(mac);
+out_put_device:
+	put_device(&dpmac_dev->dev);
 	return err;
 }
 
-- 
2.25.1

Re: [PATCH net v2 1/3] bus: fsl-mc: Fix potential double device reference in fsl_mc_get_endpoint()

[Ioana Ciornei]

On Thu, Jul 17, 2025 at 10:23:07AM +0800, Ma Ke wrote:
> The fsl_mc_get_endpoint() function may call fsl_mc_device_lookup() 
> twice, which would increment the device's reference count twice if 
> both lookups find a device. This could lead to a reference count leak.
> 
> Found by code review.
> 
> Cc: stable@vger.kernel.org
> Fixes: 1ac210d128ef ("bus: fsl-mc: add the fsl_mc_get_endpoint function")
> Signed-off-by: Ma Ke <make24@iscas.ac.cn>

Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>

Re: [PATCH net v2 3/3] dpaa2-switch: Fix device reference count leak in MAC endpoint handling

[Ioana Ciornei]

On Thu, Jul 17, 2025 at 10:23:09AM +0800, Ma Ke wrote:
> The fsl_mc_get_endpoint() function uses device_find_child() for
> localization, which implicitly calls get_device() to increment the
> device's reference count before returning the pointer. However, the
> caller dpaa2_switch_port_connect_mac() fails to properly release this 
> reference in multiple scenarios. We should call put_device() to 
> decrement reference count properly.
> 
> As comment of device_find_child() says, 'NOTE: you will need to drop
> the reference with put_device() after use'.
> 
> Found by code review.
> 
> Cc: stable@vger.kernel.org
> Fixes: 84cba72956fd ("dpaa2-switch: integrate the MAC endpoint support")
> Signed-off-by: Ma Ke <make24@iscas.ac.cn>

Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>

Re: [PATCH net v2 2/3] dpaa2-eth: Fix device reference count leak in MAC endpoint handling

[Ioana Ciornei]

On Thu, Jul 17, 2025 at 10:23:08AM +0800, Ma Ke wrote:
> The fsl_mc_get_endpoint() function uses device_find_child() for
> localization, which implicitly calls get_device() to increment the
> device's reference count before returning the pointer. However, the
> caller dpaa2_eth_connect_mac() fails to properly release this 
> reference in multiple scenarios. We should call put_device() to 
> decrement reference count properly.
> 
> As comment of device_find_child() says, 'NOTE: you will need to drop
> the reference with put_device() after use'.
> 
> Found by code review.
> 
> Cc: stable@vger.kernel.org
> Fixes: 719479230893 ("dpaa2-eth: add MAC/PHY support through phylink")
> Signed-off-by: Ma Ke <make24@iscas.ac.cn>

Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>

Re: [PATCH net v2 1/3] bus: fsl-mc: Fix potential double device reference in fsl_mc_get_endpoint()

[Simon Horman]

On Thu, Jul 17, 2025 at 10:23:07AM +0800, Ma Ke wrote:
> The fsl_mc_get_endpoint() function may call fsl_mc_device_lookup() 
> twice, which would increment the device's reference count twice if 
> both lookups find a device. This could lead to a reference count leak.
> 
> Found by code review.
> 
> Cc: stable@vger.kernel.org
> Fixes: 1ac210d128ef ("bus: fsl-mc: add the fsl_mc_get_endpoint function")

I think this should be:

Fixes: 8567494cebe5 ("bus: fsl-mc: rescan devices if endpoint not found")

I've CCed Laurentiu, the author of that commit.

> Signed-off-by: Ma Ke <make24@iscas.ac.cn>

Suggested-by: Simon Horman <horms@kernel.org>
Reviewed-by: Simon Horman <horms@kernel.org>

> ---
>  drivers/bus/fsl-mc/fsl-mc-bus.c | 19 +++++++++----------
>  1 file changed, 9 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/bus/fsl-mc/fsl-mc-bus.c b/drivers/bus/fsl-mc/fsl-mc-bus.c
> index 7671bd158545..c1c0a4759c7e 100644
> --- a/drivers/bus/fsl-mc/fsl-mc-bus.c
> +++ b/drivers/bus/fsl-mc/fsl-mc-bus.c
> @@ -943,6 +943,7 @@ struct fsl_mc_device *fsl_mc_get_endpoint(struct fsl_mc_device *mc_dev,
>  	struct fsl_mc_obj_desc endpoint_desc = {{ 0 }};
>  	struct dprc_endpoint endpoint1 = {{ 0 }};
>  	struct dprc_endpoint endpoint2 = {{ 0 }};
> +	struct fsl_mc_bus *mc_bus;
>  	int state, err;
>  
>  	mc_bus_dev = to_fsl_mc_device(mc_dev->dev.parent);
> @@ -966,6 +967,8 @@ struct fsl_mc_device *fsl_mc_get_endpoint(struct fsl_mc_device *mc_dev,
>  	strcpy(endpoint_desc.type, endpoint2.type);
>  	endpoint_desc.id = endpoint2.id;
>  	endpoint = fsl_mc_device_lookup(&endpoint_desc, mc_bus_dev);
> +	if (endpoint)
> +		return endpoint;
>  
>  	/*
>  	 * We know that the device has an endpoint because we verified by
> @@ -973,17 +976,13 @@ struct fsl_mc_device *fsl_mc_get_endpoint(struct fsl_mc_device *mc_dev,
>  	 * yet discovered by the fsl-mc bus, thus the lookup returned NULL.
>  	 * Force a rescan of the devices in this container and retry the lookup.
>  	 */
> -	if (!endpoint) {
> -		struct fsl_mc_bus *mc_bus = to_fsl_mc_bus(mc_bus_dev);
> -
> -		if (mutex_trylock(&mc_bus->scan_mutex)) {
> -			err = dprc_scan_objects(mc_bus_dev, true);
> -			mutex_unlock(&mc_bus->scan_mutex);
> -		}
> -
> -		if (err < 0)
> -			return ERR_PTR(err);
> +	mc_bus = to_fsl_mc_bus(mc_bus_dev);
> +	if (mutex_trylock(&mc_bus->scan_mutex)) {
> +		err = dprc_scan_objects(mc_bus_dev, true);
> +		mutex_unlock(&mc_bus->scan_mutex);
>  	}
> +	if (err < 0)
> +		return ERR_PTR(err);
>  
>  	endpoint = fsl_mc_device_lookup(&endpoint_desc, mc_bus_dev);
>  	/*
> -- 
> 2.25.1
> 

Re: [PATCH net v2 2/3] dpaa2-eth: Fix device reference count leak in MAC endpoint handling

[Simon Horman]

On Thu, Jul 17, 2025 at 10:23:08AM +0800, Ma Ke wrote:
> The fsl_mc_get_endpoint() function uses device_find_child() for
> localization, which implicitly calls get_device() to increment the
> device's reference count before returning the pointer. However, the
> caller dpaa2_eth_connect_mac() fails to properly release this 
> reference in multiple scenarios. We should call put_device() to 
> decrement reference count properly.
> 
> As comment of device_find_child() says, 'NOTE: you will need to drop
> the reference with put_device() after use'.
> 
> Found by code review.
> 
> Cc: stable@vger.kernel.org
> Fixes: 719479230893 ("dpaa2-eth: add MAC/PHY support through phylink")
> Signed-off-by: Ma Ke <make24@iscas.ac.cn>

Reviewed-by: Simon Horman <horms@kernel.org>

Re: [PATCH net v2 3/3] dpaa2-switch: Fix device reference count leak in MAC endpoint handling

[Simon Horman]

On Thu, Jul 17, 2025 at 10:23:09AM +0800, Ma Ke wrote:
> The fsl_mc_get_endpoint() function uses device_find_child() for
> localization, which implicitly calls get_device() to increment the
> device's reference count before returning the pointer. However, the
> caller dpaa2_switch_port_connect_mac() fails to properly release this 
> reference in multiple scenarios. We should call put_device() to 
> decrement reference count properly.
> 
> As comment of device_find_child() says, 'NOTE: you will need to drop
> the reference with put_device() after use'.
> 
> Found by code review.
> 
> Cc: stable@vger.kernel.org
> Fixes: 84cba72956fd ("dpaa2-switch: integrate the MAC endpoint support")
> Signed-off-by: Ma Ke <make24@iscas.ac.cn>

Reviewed-by: Simon Horman <horms@kernel.org>

Re: [PATCH net v2 1/3] bus: fsl-mc: Fix potential double device reference in fsl_mc_get_endpoint()

[patchwork-bot+netdevbpf]

Hello:

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

On Thu, 17 Jul 2025 10:23:07 +0800 you wrote:
> The fsl_mc_get_endpoint() function may call fsl_mc_device_lookup()
> twice, which would increment the device's reference count twice if
> both lookups find a device. This could lead to a reference count leak.
> 
> Found by code review.
> 
> Cc: stable@vger.kernel.org
> Fixes: 1ac210d128ef ("bus: fsl-mc: add the fsl_mc_get_endpoint function")
> Signed-off-by: Ma Ke <make24@iscas.ac.cn>
> 
> [...]

Here is the summary with links:
  - [net,v2,1/3] bus: fsl-mc: Fix potential double device reference in fsl_mc_get_endpoint()
    https://git.kernel.org/netdev/net/c/bddbe13d36a0
  - [net,v2,2/3] dpaa2-eth: Fix device reference count leak in MAC endpoint handling
    https://git.kernel.org/netdev/net/c/ee9f3a81ab08
  - [net,v2,3/3] dpaa2-switch: Fix device reference count leak in MAC endpoint handling
    https://git.kernel.org/netdev/net/c/96e056ffba91

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