Skip to contentSkip to conversation

net: ethernet: ti: netcp_core: fix device_node refcount leak in netcp_probe()

PATCH In review manushprajwal 2 participants 2 messages Aug 14

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 net: ethernet: ti: netcp_core: fix device_node refcount leak in netcp_probe()

From: Manush Prajwal <[email protected]>

When netcp_create_interface() failed for a given child node, netcp_probe() jumped to probe_quit_interface, which released the "interfaces" parent node but never the "child" node that the for_each_available_child_of_node() loop was iterating on.

Rework the loop around for_each_available_child_of_node_scoped() instead of adding a manual of_node_put(). child is only read inside the loop body and is never used after it, so the scoped macro's automatic cleanup on the goto is sufficient on its own.

---
v2: Rework using for_each_available_child_of_node_scoped() instead of
a manual of_node_put(), per Andrew Lunn's review.

 drivers/net/ethernet/ti/netcp_core.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/ethernet/ti/netcp_core.c +2 -2
index eb8fc2ed05..a888aa35fa 100644
@@ -2204,7 +2204,7 @@ static int netcp_probe(struct platform_device *pdev)
 {
 	struct device_node *node = pdev->dev.of_node;
 	struct netcp_intf *netcp_intf, *netcp_tmp;
-	struct device_node *child, *interfaces;
+	struct device_node *interfaces;
 	struct netcp_device *netcp_device;
 	struct device *dev = &pdev->dev;
 	struct netcp_module *module;
@@ -2246,7 +2246,7 @@ static int netcp_probe(struct platform_device *pdev)
 		goto probe_quit;
 	}
 
-	for_each_available_child_of_node(interfaces, child) {
+	for_each_available_child_of_node_scoped(interfaces, child) {
 		ret = netcp_create_interface(netcp_device, child);
 		if (ret) {
 			dev_err(dev, "could not create interface(%pOFn)\n",

--
2.46.2.windows.1

#1 lore raw
Quoted 25 lines · manushprajwal

From: Manush Prajwal <[email protected]>

When netcp_create_interface() failed for a given child node, netcp_probe() jumped to probe_quit_interface, which released the "interfaces" parent node but never the "child" node that the for_each_available_child_of_node() loop was iterating on.

Rework the loop around for_each_available_child_of_node_scoped() instead of adding a manual of_node_put(). child is only read inside the loop body and is never used after it, so the scoped macro's automatic cleanup on the goto is sufficient on its own.

---
v2: Rework using for_each_available_child_of_node_scoped() instead of
a manual of_node_put(), per Andrew Lunn's review.

 drivers/net/ethernet/ti/netcp_core.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/ethernet/ti/netcp_core.c +0 -0
index eb8fc2ed05..a888aa35fa 100644

drivers/net/ethernet/ti/netcp_core.c static int netcp_probe(struct platform_device *pdev) · line 2208

Quoted 6 lines
drivers/net/ethernet/ti/netcp_core.c +1 -1
@@ -2204,7 +2204,7 @@ static int netcp_probe(struct platform_device *pdev)
 {
 	struct device_node *node = pdev->dev.of_node;
 	struct netcp_intf *netcp_intf, *netcp_tmp;
-	struct device_node *child, *interfaces;
+	struct device_node *interfaces;
 	struct netcp_device *netcp_device;

Reverse Christmas Tree

Please also read:

https://www.kernel.org/doc/html/latest/process/maintainer-netdev.html

Andrew

---
pw-bot: cr