Skip to contentSkip to conversation

net: ibm: emac: Fix use-after-free during device removal

PATCH In review Rosen Penev 2 participants 2 messages Jun 3 – Jun 4

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

PATCH 1/1 net: ibm: emac: Fix use-after-free during device removal

The driver was using devm_register_netdev() which causes unregister_netdev() to be deferred until the devres cleanup phase, which runs after emac_remove() returns. This creates a use-after-free window where:

1. emac_remove() is called, which tears down hardware (cancels work, detaches modules, unregisters from MAL)
2. emac_remove() returns
3. devres cleanup runs and finally calls unregister_netdev()

During step 3, the network stack might still process packets, triggering emac_irq(), emac_poll(), or other handlers that access now-freed hardware resources (dev->emacp, dev->mal, etc.).

Fix this by replacing devm_register_netdev() with manual register_netdev() and calling unregister_netdev() at the beginning of emac_remove(), before any hardware teardown. This ensures the network device is fully stopped and unregistered before hardware resources are released.

The change is safe because:
- dev->ndev is assigned very early in probe (before any error paths that could bypass emac_remove)
- platform_set_drvdata() is only called after successful registration, so emac_remove() only runs for fully registered devices
- unregister_netdev() is idempotent and safe to call on any registered device

Fixes a4dd8535a527 ("net: ibm: emac: use devm for register_netdev")
Assisted-by opencode:big-pickle
---
 drivers/net/ethernet/ibm/emac/core.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/net/ethernet/ibm/emac/core.c +8 -1
index d9bbcfbcf60e..00a36c839d82 100644
@@ -3151,7 +3151,7 @@ static int emac_probe(struct platform_device *ofdev)
 
 	netif_carrier_off(ndev);
 
-	err = devm_register_netdev(&ofdev->dev, ndev);
+	err = register_netdev(ndev);
 	if (err) {
 		printk(KERN_ERR "%pOF: failed to register net device (%d)!\n",
 		       np, err);
@@ -3204,6 +3204,13 @@ static void emac_remove(struct platform_device *ofdev)
 
 	DBG(dev, "remove" NL);
 
+	/* Unregister network device before tearing down hardware
+	 * to prevent use-after-free during deferred cleanup. This ensures
+	 * the network stack stops all operations before hardware resources
+	 * are released.
+	 */
+	unregister_netdev(dev->ndev);
+
 	cancel_work_sync(&dev->reset_work);
 
 	if (emac_has_feature(dev, EMAC_FTR_HAS_TAH))

--
2.54.0

Reviewed-by on v1 1/1
#1 lore raw
Quoted 37 lines · Rosen Penev

The driver was using devm_register_netdev() which causes unregister_netdev() to be deferred until the devres cleanup phase, which runs after emac_remove() returns. This creates a use-after-free window where:

1. emac_remove() is called, which tears down hardware (cancels work, detaches modules, unregisters from MAL)
2. emac_remove() returns
3. devres cleanup runs and finally calls unregister_netdev()

During step 3, the network stack might still process packets, triggering emac_irq(), emac_poll(), or other handlers that access now-freed hardware resources (dev->emacp, dev->mal, etc.).

Fix this by replacing devm_register_netdev() with manual register_netdev() and calling unregister_netdev() at the beginning of emac_remove(), before any hardware teardown. This ensures the network device is fully stopped and unregistered before hardware resources are released.

The change is safe because:
- dev->ndev is assigned very early in probe (before any error paths that could bypass emac_remove)
- platform_set_drvdata() is only called after successful registration, so emac_remove() only runs for fully registered devices
- unregister_netdev() is idempotent and safe to call on any registered device

Fixes a4dd8535a527 ("net: ibm: emac: use devm for register_netdev")
Assisted-by opencode:big-pickle
---
 drivers/net/ethernet/ibm/emac/core.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/net/ethernet/ibm/emac/core.c +0 -0
index d9bbcfbcf60e..00a36c839d82 100644

drivers/net/ethernet/ibm/emac/core.c static int emac_probe(struct platform_device *ofdev) · line 3155

Quoted 6 lines
drivers/net/ethernet/ibm/emac/core.c +1 -1
@@ -3151,7 +3151,7 @@ static int emac_probe(struct platform_device *ofdev)
 
 	netif_carrier_off(ndev);
 
-	err = devm_register_netdev(&ofdev->dev, ndev);
+	err = register_netdev(ndev);
 	if (err) {

Right. The way devm_register_netdev *would* be safe is if everything that depends on the netdev being registered also be a devm action (since the devm cleanup actions get executed in reverse order).

Since a bunch of stuff that the netdev depends on is cleaned up normally in emac_probe, things break.

Makes sense.

Quoted 20 lines
drivers/net/ethernet/ibm/emac/core.c +7 -1
 		printk(KERN_ERR "%pOF: failed to register net device (%d)!\n",
 		       np, err);
@@ -3204,6 +3204,13 @@ static void emac_remove(struct platform_device *ofdev)
 
 	DBG(dev, "remove" NL);
 
+	/* Unregister network device before tearing down hardware
+	 * to prevent use-after-free during deferred cleanup. This ensures
+	 * the network stack stops all operations before hardware resources
+	 * are released.
+	 */
+	unregister_netdev(dev->ndev);
+
 	cancel_work_sync(&dev->reset_work);
 
 	if (emac_has_feature(dev, EMAC_FTR_HAS_TAH))
--

2.54.0