[PATCH RFC] Bluetooth: hci_sysfs: fix null pointer dereference in device_move()

[PATCH RFC] Bluetooth: hci_sysfs: fix null pointer dereference in device_move()

[syzbot]

A NULL pointer dereference in klist_put() occurs when hci_conn_del_sysfs()
attempts to reparent child devices to NULL while a child device (such as a
BNEP network device) is concurrently being unregistered:

Oops: general protection fault, probably for non-canonical address
0xdffffc000000000b: 0000 [#1] SMP KASAN NOPTI
KASAN: null-ptr-deref in range [0x0000000000000058-0x000000000000005f]
CPU: 1 UID: 0 PID: 7415 Comm: task Not tainted PREEMPT(full)
RIP: 0010:klist_put lib/klist.c:212 [inline]
RIP: 0010:klist_del lib/klist.c:230 [inline]
RIP: 0010:klist_remove+0x156/0x340 lib/klist.c:249
Call Trace:
 <TASK>
 device_move+0x18e/0x720 drivers/base/core.c:4702
 hci_conn_del_sysfs+0xb8/0x1a0 net/bluetooth/hci_sysfs.c:75
 hci_conn_cleanup net/bluetooth/hci_conn.c:170 [inline]
 hci_conn_del+0xc3d/0x1200 net/bluetooth/hci_conn.c:1318
 hci_conn_hash_flush+0x189/0x260 net/bluetooth/hci_conn.c:2747
 hci_dev_close_sync+0x7fc/0x10c0 net/bluetooth/hci_sync.c:5593
 hci_dev_do_close net/bluetooth/hci_core.c:502 [inline]
 hci_unregister_dev+0x232/0x5b0 net/bluetooth/hci_core.c:2681
 vhci_release+0x16d/0x1c0 drivers/bluetooth/hci_vhci.c:700
 ...
 </TASK>

This crash is caused by a race condition between hci_conn_del_sysfs() and
concurrent child device unregistration (e.g. bnep_session calling
unregister_netdev()). When a connection is cleaned up, hci_conn_del_sysfs()
uses device_find_any_child() to look up child devices and calls
device_move() to reparent them to NULL. If the child device is concurrently
deleted, device_del() removes the node from its parent's klist via
klist_del(&dev->p->knode_parent), which decrements the reference count and
clears the node's n_klist pointer to NULL. When device_move() subsequently
calls klist_remove(&dev->p->knode_parent), klist_put() attempts to
dereference knode_klist(n)->put, resulting in a NULL pointer dereference.

This race was introduced by commit 27aabf27fd01 ("Bluetooth: fix
use-after-free in device_for_each_child()"), which replaced
device_find_child(..., __match_tty) with device_find_any_child() in
hci_conn_del_sysfs(). That change was intended to avoid a use-after-free
where conn->dev outlived its parent hdev->dev when child devices held
references to conn->dev, because conn->dev only held a reference to
hdev->dev while registered in sysfs.

Fix the issue properly by taking an explicit reference to the parent device
with get_device(&hdev->dev) in hci_conn_init_sysfs() and dropping it with
put_device(parent) in bt_link_release() when the conn device is freed. This
ensures that hdev->dev remains valid for the entire lifecycle of conn->dev,
resolving the underlying use-after-free. With the parent reference held
properly, restore the __match_tty filter in hci_conn_del_sysfs() so that
device_move() is only invoked on persistent RFCOMM TTY devices as
originally intended, eliminating the race condition with unregistering
network devices.

Fixes: 27aabf27fd01 ("Bluetooth: fix use-after-free in device_for_each_child()")
Assisted-by: Gemini:gemini-3.7-flash Gemini:gemini-3.1-pro-preview syzbot
Reported-by: syzbot+6df45dd3d03e1a9aca96@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=6df45dd3d03e1a9aca96
Link: https://syzkaller.appspot.com/ai_job?id=2d312ca9-20bf-457f-8a4f-728831a400af
To: <linux-bluetooth@vger.kernel.org>
To: "Luiz Augusto von Dentz" <luiz.dentz@gmail.com>
To: "Marcel Holtmann" <marcel@holtmann.org>
To: "Dmitry Antipov" <dmantipov@yandex.ru>
Cc: <linux-kernel@vger.kernel.org>

---
diff --git a/net/bluetooth/hci_sysfs.c b/net/bluetooth/hci_sysfs.c
index 8957ce7c2..c2065abf7 100644
--- a/net/bluetooth/hci_sysfs.c
+++ b/net/bluetooth/hci_sysfs.c
@@ -13,7 +13,10 @@ static const struct class bt_class = {
 static void bt_link_release(struct device *dev)
 {
 	struct hci_conn *conn = to_hci_conn(dev);
+	struct device *parent = dev->parent;
+
 	kfree(conn);
+	put_device(parent);
 }
 
 static const struct device_type bt_link = {
@@ -21,6 +24,16 @@ static const struct device_type bt_link = {
 	.release = bt_link_release,
 };
 
+/*
+ * The rfcomm tty device will possibly retain even when conn
+ * is down, and sysfs doesn't support move zombie device,
+ * so we should move the device before conn device is destroyed.
+ */
+static int __match_tty(struct device *dev, const void *data)
+{
+	return !strncmp(dev_name(dev), "rfcomm", 6);
+}
+
 void hci_conn_init_sysfs(struct hci_conn *conn)
 {
 	struct hci_dev *hdev = conn->hdev;
@@ -29,7 +42,7 @@ void hci_conn_init_sysfs(struct hci_conn *conn)
 
 	conn->dev.type = &bt_link;
 	conn->dev.class = &bt_class;
-	conn->dev.parent = &hdev->dev;
+	conn->dev.parent = get_device(&hdev->dev);
 
 	device_initialize(&conn->dev);
 }
@@ -69,7 +82,7 @@ void hci_conn_del_sysfs(struct hci_conn *conn)
 	while (1) {
 		struct device *dev;
 
-		dev = device_find_any_child(&conn->dev);
+		dev = device_find_child(&conn->dev, NULL, __match_tty);
 		if (!dev)
 			break;
 		device_move(dev, NULL, DPM_ORDER_DEV_LAST);


base-commit: cee9395acd8043be0644b25c34bfa86623f2b935
-- 
This is an AI-generated patch subject to moderation.
Reply with '#syz upstream' to Sign-off the patch as a human author
and send it to the upstream kernel mailing lists.
Reply with '#syz reject' to reject it ('#syz unreject' to undo).

See https://goo.gle/syzbot-ai-patches for information about AI-generated patches.
You can comment on the patch as usual, syzbot will try to address
the comments and send a new version of the patch if necessary.
syzbot engineers can be reached at syzkaller@googlegroups.com.

Re: [PATCH RFC] Bluetooth: hci_sysfs: fix null pointer dereference in device_move()

[Krystian Kaniewski]

Correct the subject and commit message to match the retained crash 
evidence. The report faults in klist_put() from device_del() during 
bnep_session, not in device_move(), and it does not contain the quoted 
PID 7415 "Comm: task" trace. Remove that unsupported trace or replace it 
with the actual bnep_session and device_del stack. Describe the observed 
ordering: device_del() snapshots a non-NULL parent, concurrent 
hci_conn_del_sysfs() moves the child and clears knode_parent, then 
device_del() calls klist_del() using its stale parent decision. Preserve 
the code diff, which correctly restores the RFCOMM-only move and retains 
hdev->dev until the connection device release, along with the existing 
Fixes, Reported-by, Closes, provenance, and recipient tags.

On 9/1/2026 12:16 AM, syzbot wrote:
> A NULL pointer dereference in klist_put() occurs when hci_conn_del_sysfs()
> attempts to reparent child devices to NULL while a child device (such as a
> BNEP network device) is concurrently being unregistered:
>
> Oops: general protection fault, probably for non-canonical address
> 0xdffffc000000000b: 0000 [#1] SMP KASAN NOPTI
> KASAN: null-ptr-deref in range [0x0000000000000058-0x000000000000005f]
> CPU: 1 UID: 0 PID: 7415 Comm: task Not tainted PREEMPT(full)
> RIP: 0010:klist_put lib/klist.c:212 [inline]
> RIP: 0010:klist_del lib/klist.c:230 [inline]
> RIP: 0010:klist_remove+0x156/0x340 lib/klist.c:249
> Call Trace:
>   <TASK>
>   device_move+0x18e/0x720 drivers/base/core.c:4702
>   hci_conn_del_sysfs+0xb8/0x1a0 net/bluetooth/hci_sysfs.c:75
>   hci_conn_cleanup net/bluetooth/hci_conn.c:170 [inline]
>   hci_conn_del+0xc3d/0x1200 net/bluetooth/hci_conn.c:1318
>   hci_conn_hash_flush+0x189/0x260 net/bluetooth/hci_conn.c:2747
>   hci_dev_close_sync+0x7fc/0x10c0 net/bluetooth/hci_sync.c:5593
>   hci_dev_do_close net/bluetooth/hci_core.c:502 [inline]
>   hci_unregister_dev+0x232/0x5b0 net/bluetooth/hci_core.c:2681
>   vhci_release+0x16d/0x1c0 drivers/bluetooth/hci_vhci.c:700
>   ...
>   </TASK>
>
> This crash is caused by a race condition between hci_conn_del_sysfs() and
> concurrent child device unregistration (e.g. bnep_session calling
> unregister_netdev()). When a connection is cleaned up, hci_conn_del_sysfs()
> uses device_find_any_child() to look up child devices and calls
> device_move() to reparent them to NULL. If the child device is concurrently
> deleted, device_del() removes the node from its parent's klist via
> klist_del(&dev->p->knode_parent), which decrements the reference count and
> clears the node's n_klist pointer to NULL. When device_move() subsequently
> calls klist_remove(&dev->p->knode_parent), klist_put() attempts to
> dereference knode_klist(n)->put, resulting in a NULL pointer dereference.
>
> This race was introduced by commit 27aabf27fd01 ("Bluetooth: fix
> use-after-free in device_for_each_child()"), which replaced
> device_find_child(..., __match_tty) with device_find_any_child() in
> hci_conn_del_sysfs(). That change was intended to avoid a use-after-free
> where conn->dev outlived its parent hdev->dev when child devices held
> references to conn->dev, because conn->dev only held a reference to
> hdev->dev while registered in sysfs.
>
> Fix the issue properly by taking an explicit reference to the parent device
> with get_device(&hdev->dev) in hci_conn_init_sysfs() and dropping it with
> put_device(parent) in bt_link_release() when the conn device is freed. This
> ensures that hdev->dev remains valid for the entire lifecycle of conn->dev,
> resolving the underlying use-after-free. With the parent reference held
> properly, restore the __match_tty filter in hci_conn_del_sysfs() so that
> device_move() is only invoked on persistent RFCOMM TTY devices as
> originally intended, eliminating the race condition with unregistering
> network devices.
>
> Fixes: 27aabf27fd01 ("Bluetooth: fix use-after-free in device_for_each_child()")
> Assisted-by: Gemini:gemini-3.7-flash Gemini:gemini-3.1-pro-preview syzbot
> Reported-by: syzbot+6df45dd3d03e1a9aca96@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=6df45dd3d03e1a9aca96
> Link: https://syzkaller.appspot.com/ai_job?id=2d312ca9-20bf-457f-8a4f-728831a400af
> To: <linux-bluetooth@vger.kernel.org>
> To: "Luiz Augusto von Dentz" <luiz.dentz@gmail.com>
> To: "Marcel Holtmann" <marcel@holtmann.org>
> To: "Dmitry Antipov" <dmantipov@yandex.ru>
> Cc: <linux-kernel@vger.kernel.org>
>
> ---
> diff --git a/net/bluetooth/hci_sysfs.c b/net/bluetooth/hci_sysfs.c
> index 8957ce7c2..c2065abf7 100644
> --- a/net/bluetooth/hci_sysfs.c
> +++ b/net/bluetooth/hci_sysfs.c
> @@ -13,7 +13,10 @@ static const struct class bt_class = {
>   static void bt_link_release(struct device *dev)
>   {
>   	struct hci_conn *conn = to_hci_conn(dev);
> +	struct device *parent = dev->parent;
> +
>   	kfree(conn);
> +	put_device(parent);
>   }
>   
>   static const struct device_type bt_link = {
> @@ -21,6 +24,16 @@ static const struct device_type bt_link = {
>   	.release = bt_link_release,
>   };
>   
> +/*
> + * The rfcomm tty device will possibly retain even when conn
> + * is down, and sysfs doesn't support move zombie device,
> + * so we should move the device before conn device is destroyed.
> + */
> +static int __match_tty(struct device *dev, const void *data)
> +{
> +	return !strncmp(dev_name(dev), "rfcomm", 6);
> +}
> +
>   void hci_conn_init_sysfs(struct hci_conn *conn)
>   {
>   	struct hci_dev *hdev = conn->hdev;
> @@ -29,7 +42,7 @@ void hci_conn_init_sysfs(struct hci_conn *conn)
>   
>   	conn->dev.type = &bt_link;
>   	conn->dev.class = &bt_class;
> -	conn->dev.parent = &hdev->dev;
> +	conn->dev.parent = get_device(&hdev->dev);
>   
>   	device_initialize(&conn->dev);
>   }
> @@ -69,7 +82,7 @@ void hci_conn_del_sysfs(struct hci_conn *conn)
>   	while (1) {
>   		struct device *dev;
>   
> -		dev = device_find_any_child(&conn->dev);
> +		dev = device_find_child(&conn->dev, NULL, __match_tty);
>   		if (!dev)
>   			break;
>   		device_move(dev, NULL, DPM_ORDER_DEV_LAST);
>
>
> base-commit: cee9395acd8043be0644b25c34bfa86623f2b935