linux-kernel - [PATCH v2] Bluetooth: skip invalid hci_sync_conn_complete

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20210728075105.415214-1-desmondcheongzx@gmail.com>
Date:   Wed, 28 Jul 2021 15:51:04 +0800
From:   Desmond Cheong Zhi Xi <desmondcheongzx@...il.com>
To:     marcel@...tmann.org, johan.hedberg@...il.com, luiz.dentz@...il.com,
        davem@...emloft.net, kuba@...nel.org
Cc:     Desmond Cheong Zhi Xi <desmondcheongzx@...il.com>,
        linux-bluetooth@...r.kernel.org, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org, skhan@...uxfoundation.org,
        gregkh@...uxfoundation.org,
        linux-kernel-mentees@...ts.linuxfoundation.org,
        syzbot+66264bf2fd0476be7e6c@...kaller.appspotmail.com
Subject: [PATCH v2] Bluetooth: skip invalid hci_sync_conn_complete_evt

Syzbot reported a corrupted list in kobject_add_internal [1]. This
happens when multiple HCI_EV_SYNC_CONN_COMPLETE event packets with
status 0 are sent for the same HCI connection. This causes us to
register the device more than once which corrupts the kset list.

As this is forbidden behavior, we add a check for whether we're
trying to process the same HCI_EV_SYNC_CONN_COMPLETE event multiple
times for one connection. If that's the case, the event is invalid, so
we report an error that the device is misbehaving, and ignore the
packet.

Link: https://syzkaller.appspot.com/bug?extid=66264bf2fd0476be7e6c [1]
Reported-by: syzbot+66264bf2fd0476be7e6c@...kaller.appspotmail.com
Tested-by: syzbot+66264bf2fd0476be7e6c@...kaller.appspotmail.com
Signed-off-by: Desmond Cheong Zhi Xi <desmondcheongzx@...il.com>
---

v1 -> v2:
- Added more comments to explain the reasoning behind the new check, and
a bt_dev_err message upon detecting the invalid event. As suggested by
Marcel Holtmann.

 net/bluetooth/hci_event.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
index 016b2999f219..a6df4f9d2c23 100644
--- a/net/bluetooth/hci_event.c
+++ b/net/bluetooth/hci_event.c
@@ -4373,6 +4373,22 @@ static void hci_sync_conn_complete_evt(struct hci_dev *hdev,
 
 	switch (ev->status) {
 	case 0x00:
+		/* The synchronous connection complete event should only be
+		 * sent once per new connection. Receiving a successful
+		 * complete event when the connection status is already
+		 * BT_CONNECTED means that the device is misbehaving and sent
+		 * multiple complete event packets for the same new connection.
+		 *
+		 * Registering the device more than once can corrupt kernel
+		 * memory, hence upon detecting this invalid event, we report
+		 * an error and ignore the packet.
+		 */
+		if (conn->state == BT_CONNECTED) {
+			bt_dev_err(hdev,
+				   "received multiple HCI_EV_SYNC_CONN_COMPLETE events with status 0 for conn %p",
+				   conn);
+			goto unlock;
+		}
 		conn->handle = __le16_to_cpu(ev->handle);
 		conn->state  = BT_CONNECTED;
 		conn->type   = ev->link_type;
-- 
2.25.1