[PATCH] net: caif: serial: fix TX UAF on ser->tty

[PATCH] net: caif: serial: fix TX UAF on ser->tty

[Shuangpeng Bai]

KASAN reported a slab-use-after-free in tty_write_room() reachable from
caif_serial's TX path. The TX handler dereferences ser->tty while
ldisc_close() can drop the driver's tty reference. Since ser->tty was
not cleared and accesses were not synchronized, the TX path could race
with tty teardown and dereference a stale ser->tty pointer.

Fix it by serializing accesses to ser->tty with a dedicated lock. The TX
path grabs a tty kref under the lock and drops it after the TX attempt,
while ldisc_close() clears ser->tty under the same lock before putting
the old tty reference. This prevents the TX path from observing a freed
tty object via ser->tty.

Reported-by: Shuangpeng Bai <baisp@psu.edu>
Closes: https://groups.google.com/g/syzkaller/c/usNe0oKtoXw/m/x8qUc3yUAQAJ

Signed-off-by: Shuangpeng Bai <shuangpeng.kernel@gmail.com>
---
 drivers/net/caif/caif_serial.c | 22 ++++++++++++++++++++--
 1 file changed, 20 insertions(+), 2 deletions(-)

diff --git a/drivers/net/caif/caif_serial.c b/drivers/net/caif/caif_serial.c
index c398ac42eae9..fd1213685a89 100644
--- a/drivers/net/caif/caif_serial.c
+++ b/drivers/net/caif/caif_serial.c
@@ -68,6 +68,7 @@ struct ser_device {
 	struct net_device *dev;
 	struct sk_buff_head head;
 	struct tty_struct *tty;
+	spinlock_t tty_lock; /* protects ser->tty */
 	bool tx_started;
 	unsigned long state;
 #ifdef CONFIG_DEBUG_FS
@@ -197,12 +198,21 @@ static int handle_tx(struct ser_device *ser)
 	struct sk_buff *skb;
 	int tty_wr, len, room;
 
+	spin_lock(&ser->tty_lock);
 	tty = ser->tty;
+	tty_kref_get(tty);
+	spin_unlock(&ser->tty_lock);
+
+	if (!tty)
+		return 0;
+
 	ser->tx_started = true;
 
 	/* Enter critical section */
-	if (test_and_set_bit(CAIF_SENDING, &ser->state))
+	if (test_and_set_bit(CAIF_SENDING, &ser->state)) {
+		tty_kref_put(tty);
 		return 0;
+	}
 
 	/* skb_peek is safe because handle_tx is called after skb_queue_tail */
 	while ((skb = skb_peek(&ser->head)) != NULL) {
@@ -245,9 +255,11 @@ static int handle_tx(struct ser_device *ser)
 		ser->common.flowctrl != NULL)
 				ser->common.flowctrl(ser->dev, ON);
 	clear_bit(CAIF_SENDING, &ser->state);
+	tty_kref_put(tty);
 	return 0;
 error:
 	clear_bit(CAIF_SENDING, &ser->state);
+	tty_kref_put(tty);
 	return tty_wr;
 }
 
@@ -327,6 +339,7 @@ static int ldisc_open(struct tty_struct *tty)
 		return -ENOMEM;
 
 	ser = netdev_priv(dev);
+	spin_lock_init(&ser->tty_lock);
 	ser->tty = tty_kref_get(tty);
 	ser->dev = dev;
 	debugfs_init(ser, tty);
@@ -354,8 +367,13 @@ static int ldisc_open(struct tty_struct *tty)
 static void ldisc_close(struct tty_struct *tty)
 {
 	struct ser_device *ser = tty->disc_data;
+	struct tty_struct *old;
 
-	tty_kref_put(ser->tty);
+	spin_lock(&ser->tty_lock);
+	old = ser->tty;
+	ser->tty = NULL;
+	spin_unlock(&ser->tty_lock);
+	tty_kref_put(old);
 
 	spin_lock(&ser_lock);
 	list_move(&ser->node, &ser_release_list);
-- 
2.34.1

Re: [PATCH] net: caif: serial: fix TX UAF on ser->tty

[Paolo Abeni]

On 2/12/26 5:22 AM, Shuangpeng Bai wrote:
> KASAN reported a slab-use-after-free in tty_write_room() reachable from
> caif_serial's TX path. The TX handler dereferences ser->tty while
> ldisc_close() can drop the driver's tty reference. Since ser->tty was
> not cleared and accesses were not synchronized, the TX path could race
> with tty teardown and dereference a stale ser->tty pointer.
> 
> Fix it by serializing accesses to ser->tty with a dedicated lock. The TX
> path grabs a tty kref under the lock and drops it after the TX attempt,
> while ldisc_close() clears ser->tty under the same lock before putting
> the old tty reference. This prevents the TX path from observing a freed
> tty object via ser->tty.
> 
> Reported-by: Shuangpeng Bai <baisp@psu.edu>
> Closes: https://groups.google.com/g/syzkaller/c/usNe0oKtoXw/m/x8qUc3yUAQAJ
> 

Please, no empty lines in the tag area.

You must include a fixes tag and specify the target tree in the subj prefix

Does not apply cleanly to net nor net-next

Please read carefully the process documentation under:

Documentation/process/

and especially Documentation/process/maintainer-netdev.rst

before resubmitting

/P