diff mbox series

Bluetooth: L2CAP: Fix div-by-zero in l2cap_le_flowctl_init()

Message ID 20240428054307.1178347-1-iam@sung-woo.kim
State New
Headers show
Series Bluetooth: L2CAP: Fix div-by-zero in l2cap_le_flowctl_init() | expand

Commit Message

Sungwoo Kim April 28, 2024, 5:43 a.m. UTC
Hello, could you review this bug and its patch?

l2cap_le_flowctl_init() can cause both div-by-zero and an integer overflow.

l2cap_le_flowctl_init()
  chan->mps = min_t(u16, chan->imtu, chan->conn->mtu - L2CAP_HDR_SIZE);
  chan->rx_credits = (chan->imtu / chan->mps) + 1;  <- div-by-zero

Here, mtu could be less than or equal to L2CAP_HDR_SIZE (4). If mtu is 4, it
causes div-by-zero. If mtu is less than 4, it causes an integer overflow.

How mtu could have such low value:

hci_cc_le_read_buffer_size()
  hdev->le_mtu = __le16_to_cpu(rp->le_mtu);

l2cap_conn_add()
  conn->mtu = hcon->hdev->le_mtu;

As shown, mtu is an input from an HCI device. So, any HCI device can
set mtu value to any value, such as lower than 4.

To fix this, this patch adds a validation before subtractions. If MTU is
too small, the corresponding value will set by 0, and a warning message
will show up.

However, I'm not sure that 0-ing the value is the best. It'd be great if
you could comment on this.

Thank you,
Sungwoo.

divide error: 0000 [#1] PREEMPT SMP KASAN NOPTI
CPU: 0 PID: 67 Comm: kworker/u5:0 Tainted: G        W          6.9.0-rc5+ #20
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1 04/01/2014
Workqueue: hci0 hci_rx_work
RIP: 0010:l2cap_le_flowctl_init+0x19e/0x3f0 net/bluetooth/l2cap_core.c:547
Code: e8 17 17 0c 00 66 41 89 9f 84 00 00 00 bf 01 00 00 00 41 b8 02 00 00 00 4c 89 fe 4c 89 e2 89 d9 e8 27 17 0c 00 44 89 f0 31 d2 <66> f7 f3 89 c3 ff c3 4d 8d b7 88 00 00 00 4c 89 f0 48 c1 e8 03 42
RSP: 0018:ffff88810bc0f858 EFLAGS: 00010246
RAX: 00000000000002a0 RBX: 0000000000000000 RCX: dffffc0000000000
RDX: 0000000000000000 RSI: ffff88810bc0f7c0 RDI: ffffc90002dcb66f
RBP: ffff88810bc0f880 R08: aa69db2dda70ff01 R09: 0000ffaaaaaaaaaa
R10: 0084000000ffaaaa R11: 0000000000000000 R12: ffff88810d65a084
R13: dffffc0000000000 R14: 00000000000002a0 R15: ffff88810d65a000
FS:  0000000000000000(0000) GS:ffff88811ac00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000020000100 CR3: 0000000103268003 CR4: 0000000000770ef0
PKRU: 55555554
Call Trace:
 <TASK>
 l2cap_le_connect_req net/bluetooth/l2cap_core.c:4902 [inline]
 l2cap_le_sig_cmd net/bluetooth/l2cap_core.c:5420 [inline]
 l2cap_le_sig_channel net/bluetooth/l2cap_core.c:5486 [inline]
 l2cap_recv_frame+0xe59d/0x11710 net/bluetooth/l2cap_core.c:6809
 l2cap_recv_acldata+0x544/0x10a0 net/bluetooth/l2cap_core.c:7506
 hci_acldata_packet net/bluetooth/hci_core.c:3939 [inline]
 hci_rx_work+0x5e5/0xb20 net/bluetooth/hci_core.c:4176
 process_one_work kernel/workqueue.c:3254 [inline]
 process_scheduled_works+0x90f/0x1530 kernel/workqueue.c:3335
 worker_thread+0x926/0xe70 kernel/workqueue.c:3416
 kthread+0x2e3/0x380 kernel/kthread.c:388
 ret_from_fork+0x5c/0x90 arch/x86/kernel/process.c:147
 ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
 </TASK>
Modules linked in:
---[ end trace 0000000000000000 ]---

Signed-off-by: Sungwoo Kim <iam@sung-woo.kim>
---
 net/bluetooth/l2cap_core.c | 94 +++++++++++++++++++++++++++++---------
 1 file changed, 73 insertions(+), 21 deletions(-)

Comments

Sungwoo Kim May 1, 2024, 10:23 p.m. UTC | #1
Dear Luiz,

On Mon, Apr 29, 2024 at 11:15 AM Luiz Augusto von Dentz
<luiz.dentz@gmail.com> wrote:
>
> Hi Sungwoo,
>
> On Sun, Apr 28, 2024 at 1:43 AM Sungwoo Kim <iam@sung-woo.kim> wrote:
> >
> > Hello, could you review this bug and its patch?
> >
> > l2cap_le_flowctl_init() can cause both div-by-zero and an integer overflow.
> >
> > l2cap_le_flowctl_init()
> >   chan->mps = min_t(u16, chan->imtu, chan->conn->mtu - L2CAP_HDR_SIZE);
> >   chan->rx_credits = (chan->imtu / chan->mps) + 1;  <- div-by-zero
> >
> > Here, mtu could be less than or equal to L2CAP_HDR_SIZE (4). If mtu is 4, it
> > causes div-by-zero. If mtu is less than 4, it causes an integer overflow.
>
> That is because it is not valid to have hdev->le_mtu < 0x001b (the
> range is 0x001b to 0xffff), so we should really look into checking
> that conn->mtu is actually valid.
>
> > How mtu could have such low value:
> >
> > hci_cc_le_read_buffer_size()
> >   hdev->le_mtu = __le16_to_cpu(rp->le_mtu);
> >
> > l2cap_conn_add()
> >   conn->mtu = hcon->hdev->le_mtu;
>
> Yeah this assignment is incorrect and in fact we don't do that if
> le_mtu is zero so we probably should do some checks e.g. le_mtu >
> 0x001a, or perhaps we need to move the MTU directly to hci_conn so it
> can check there are enough buffers to serve the link so we stop the
> connection procedure earlier.

Let's say we moved MTU directly to hci_conn and already checked enough
buffers at the creation of hcon.
Then, what should happen if hdev->le_mtu is updated? (by a new
le_read_buffer_size cmd)
Should hcon->mtu be synced with hdev->le_mtu? Or hcon->mtu can keep
its old value?

Best,
Sungwoo.
Luiz Augusto von Dentz May 1, 2024, 10:36 p.m. UTC | #2
Hi Sungwoo,

On Wed, May 1, 2024 at 6:23 PM Sungwoo Kim <iam@sung-woo.kim> wrote:
>
> Dear Luiz,
>
> On Mon, Apr 29, 2024 at 11:15 AM Luiz Augusto von Dentz
> <luiz.dentz@gmail.com> wrote:
> >
> > Hi Sungwoo,
> >
> > On Sun, Apr 28, 2024 at 1:43 AM Sungwoo Kim <iam@sung-woo.kim> wrote:
> > >
> > > Hello, could you review this bug and its patch?
> > >
> > > l2cap_le_flowctl_init() can cause both div-by-zero and an integer overflow.
> > >
> > > l2cap_le_flowctl_init()
> > >   chan->mps = min_t(u16, chan->imtu, chan->conn->mtu - L2CAP_HDR_SIZE);
> > >   chan->rx_credits = (chan->imtu / chan->mps) + 1;  <- div-by-zero
> > >
> > > Here, mtu could be less than or equal to L2CAP_HDR_SIZE (4). If mtu is 4, it
> > > causes div-by-zero. If mtu is less than 4, it causes an integer overflow.
> >
> > That is because it is not valid to have hdev->le_mtu < 0x001b (the
> > range is 0x001b to 0xffff), so we should really look into checking
> > that conn->mtu is actually valid.
> >
> > > How mtu could have such low value:
> > >
> > > hci_cc_le_read_buffer_size()
> > >   hdev->le_mtu = __le16_to_cpu(rp->le_mtu);
> > >
> > > l2cap_conn_add()
> > >   conn->mtu = hcon->hdev->le_mtu;
> >
> > Yeah this assignment is incorrect and in fact we don't do that if
> > le_mtu is zero so we probably should do some checks e.g. le_mtu >
> > 0x001a, or perhaps we need to move the MTU directly to hci_conn so it
> > can check there are enough buffers to serve the link so we stop the
> > connection procedure earlier.
>
> Let's say we moved MTU directly to hci_conn and already checked enough
> buffers at the creation of hcon.
> Then, what should happen if hdev->le_mtu is updated? (by a new
> le_read_buffer_size cmd)
> Should hcon->mtu be synced with hdev->le_mtu? Or hcon->mtu can keep
> its old value?

What now, why would we read it again? These commands are only suppose
to be send during init phase and if you do this sort of thing as an
event without a command then it is also against the spec and we shall
probably just ignore it if there is no command pending.

> Best,
> Sungwoo.
diff mbox series

Patch

diff --git a/net/bluetooth/l2cap_core.c b/net/bluetooth/l2cap_core.c
index 84fc70862..472ddfb2e 100644
--- a/net/bluetooth/l2cap_core.c
+++ b/net/bluetooth/l2cap_core.c
@@ -541,10 +541,17 @@  static void l2cap_le_flowctl_init(struct l2cap_chan *chan, u16 tx_credits)
 	chan->sdu_last_frag = NULL;
 	chan->sdu_len = 0;
 	chan->tx_credits = tx_credits;
-	/* Derive MPS from connection MTU to stop HCI fragmentation */
-	chan->mps = min_t(u16, chan->imtu, chan->conn->mtu - L2CAP_HDR_SIZE);
-	/* Give enough credits for a full packet */
-	chan->rx_credits = (chan->imtu / chan->mps) + 1;
+
+	if (chan->conn->mtu < L2CAP_HDR_SIZE) {
+		BT_WARN("mtu is too small (mtu %d < %d)", chan->conn->mtu, L2CAP_HDR_SIZE);
+		chan->mps = 0;
+		chan->rx_credits = 0;
+	} else {
+		/* Derive MPS from connection MTU to stop HCI fragmentation */
+		chan->mps = min_t(u16, chan->imtu, chan->conn->mtu - L2CAP_HDR_SIZE);
+		/* Give enough credits for a full packet */
+		chan->rx_credits = (chan->imtu / chan->mps) + 1;
+	}
 
 	skb_queue_head_init(&chan->tx_q);
 }
@@ -2222,7 +2229,12 @@  static struct sk_buff *l2cap_create_connless_pdu(struct l2cap_chan *chan,
 	BT_DBG("chan %p psm 0x%2.2x len %zu", chan,
 	       __le16_to_cpu(chan->psm), len);
 
-	count = min_t(unsigned int, (conn->mtu - hlen), len);
+	if (conn->mtu < hlen) {
+		count = 0;
+		BT_WARN("mtu is too small (mtu %d < len %d)", conn->mtu, hlen);
+	} else {
+		count = min_t(unsigned int, (conn->mtu - hlen), len);
+	}
 
 	skb = chan->ops->alloc_skb(chan, hlen, count,
 				   msg->msg_flags & MSG_DONTWAIT);
@@ -2253,7 +2265,12 @@  static struct sk_buff *l2cap_create_basic_pdu(struct l2cap_chan *chan,
 
 	BT_DBG("chan %p len %zu", chan, len);
 
-	count = min_t(unsigned int, (conn->mtu - L2CAP_HDR_SIZE), len);
+	if (conn->mtu < L2CAP_HDR_SIZE) {
+		BT_WARN("mtu is too small (mtu %d < %d)", conn->mtu, L2CAP_HDR_SIZE);
+		count = 0;
+	} else {
+		count = min_t(unsigned int, (conn->mtu - L2CAP_HDR_SIZE), len);
+	}
 
 	skb = chan->ops->alloc_skb(chan, L2CAP_HDR_SIZE, count,
 				   msg->msg_flags & MSG_DONTWAIT);
@@ -2295,7 +2312,12 @@  static struct sk_buff *l2cap_create_iframe_pdu(struct l2cap_chan *chan,
 	if (chan->fcs == L2CAP_FCS_CRC16)
 		hlen += L2CAP_FCS_SIZE;
 
-	count = min_t(unsigned int, (conn->mtu - hlen), len);
+	if (conn->mtu < hlen) {
+		BT_WARN("mtu is too small (mtu %d < len %d)", conn->mtu, hlen);
+		count = 0;
+	} else {
+		count = min_t(unsigned int, (conn->mtu - hlen), len);
+	}
 
 	skb = chan->ops->alloc_skb(chan, hlen, count,
 				   msg->msg_flags & MSG_DONTWAIT);
@@ -2412,7 +2434,12 @@  static struct sk_buff *l2cap_create_le_flowctl_pdu(struct l2cap_chan *chan,
 	if (sdulen)
 		hlen += L2CAP_SDULEN_SIZE;
 
-	count = min_t(unsigned int, (conn->mtu - hlen), len);
+	if (conn->mtu < hlen) {
+		BT_WARN("mtu is too small (mtu %d < len %d)", conn->mtu, hlen);
+		count = 0;
+	} else {
+		count = min_t(unsigned int, (conn->mtu - hlen), len);
+	}
 
 	skb = chan->ops->alloc_skb(chan, hlen, count,
 				   msg->msg_flags & MSG_DONTWAIT);
@@ -3225,6 +3252,7 @@  static int l2cap_build_conf_req(struct l2cap_chan *chan, void *data, size_t data
 	void *ptr = req->data;
 	void *endptr = data + data_size;
 	u16 size;
+	int hlen;
 
 	BT_DBG("chan %p", chan);
 
@@ -3275,14 +3303,19 @@  static int l2cap_build_conf_req(struct l2cap_chan *chan, void *data, size_t data
 		break;
 
 	case L2CAP_MODE_ERTM:
+		hlen = L2CAP_EXT_HDR_SIZE + L2CAP_SDULEN_SIZE + L2CAP_FCS_SIZE;
 		rfc.mode            = L2CAP_MODE_ERTM;
 		rfc.max_transmit    = chan->max_tx;
 
 		__l2cap_set_ertm_timeouts(chan, &rfc);
 
-		size = min_t(u16, L2CAP_DEFAULT_MAX_PDU_SIZE, chan->conn->mtu -
-			     L2CAP_EXT_HDR_SIZE - L2CAP_SDULEN_SIZE -
-			     L2CAP_FCS_SIZE);
+		if (chan->conn->mtu < hlen) {
+			BT_WARN("mtu is too small (mtu %d < len %d)", chan->conn->mtu, hlen);
+			size = 0;
+		} else {
+			size = min_t(u16, L2CAP_DEFAULT_MAX_PDU_SIZE, chan->conn->mtu - hlen);
+		}
+
 		rfc.max_pdu_size = cpu_to_le16(size);
 
 		l2cap_txwin_setup(chan);
@@ -3310,6 +3343,7 @@  static int l2cap_build_conf_req(struct l2cap_chan *chan, void *data, size_t data
 		break;
 
 	case L2CAP_MODE_STREAMING:
+		hlen = L2CAP_EXT_HDR_SIZE + L2CAP_SDULEN_SIZE + L2CAP_FCS_SIZE;
 		l2cap_txwin_setup(chan);
 		rfc.mode            = L2CAP_MODE_STREAMING;
 		rfc.txwin_size      = 0;
@@ -3317,9 +3351,12 @@  static int l2cap_build_conf_req(struct l2cap_chan *chan, void *data, size_t data
 		rfc.retrans_timeout = 0;
 		rfc.monitor_timeout = 0;
 
-		size = min_t(u16, L2CAP_DEFAULT_MAX_PDU_SIZE, chan->conn->mtu -
-			     L2CAP_EXT_HDR_SIZE - L2CAP_SDULEN_SIZE -
-			     L2CAP_FCS_SIZE);
+		if (chan->conn->mtu < hlen) {
+			BT_WARN("mtu is too small (mtu %d < len %d)", chan->conn->mtu, hlen);
+			size = 0;
+		} else {
+			size = min_t(u16, L2CAP_DEFAULT_MAX_PDU_SIZE, chan->conn->mtu - hlen);
+		}
 		rfc.max_pdu_size = cpu_to_le16(size);
 
 		l2cap_add_conf_opt(&ptr, L2CAP_CONF_RFC, sizeof(rfc),
@@ -3351,7 +3388,7 @@  static int l2cap_parse_conf_req(struct l2cap_chan *chan, void *data, size_t data
 	void *endptr = data + data_size;
 	void *req = chan->conf_req;
 	int len = chan->conf_len;
-	int type, hint, olen;
+	int type, hint, olen, hlen;
 	unsigned long val;
 	struct l2cap_conf_rfc rfc = { .mode = L2CAP_MODE_BASIC };
 	struct l2cap_conf_efs efs;
@@ -3496,6 +3533,7 @@  static int l2cap_parse_conf_req(struct l2cap_chan *chan, void *data, size_t data
 			break;
 
 		case L2CAP_MODE_ERTM:
+			hlen = L2CAP_EXT_HDR_SIZE + L2CAP_SDULEN_SIZE + L2CAP_FCS_SIZE;
 			if (!test_bit(CONF_EWS_RECV, &chan->conf_state))
 				chan->remote_tx_win = rfc.txwin_size;
 			else
@@ -3503,9 +3541,15 @@  static int l2cap_parse_conf_req(struct l2cap_chan *chan, void *data, size_t data
 
 			chan->remote_max_tx = rfc.max_transmit;
 
-			size = min_t(u16, le16_to_cpu(rfc.max_pdu_size),
-				     chan->conn->mtu - L2CAP_EXT_HDR_SIZE -
-				     L2CAP_SDULEN_SIZE - L2CAP_FCS_SIZE);
+			if (chan->conn->mtu < hlen) {
+				BT_WARN("mtu is too small (mtu %d < len %d)",
+					chan->conn->mtu, hlen);
+				size = 0;
+			} else {
+				size = min_t(u16, le16_to_cpu(rfc.max_pdu_size),
+					     chan->conn->mtu - hlen);
+			}
+
 			rfc.max_pdu_size = cpu_to_le16(size);
 			chan->remote_mps = size;
 
@@ -3534,9 +3578,17 @@  static int l2cap_parse_conf_req(struct l2cap_chan *chan, void *data, size_t data
 			break;
 
 		case L2CAP_MODE_STREAMING:
-			size = min_t(u16, le16_to_cpu(rfc.max_pdu_size),
-				     chan->conn->mtu - L2CAP_EXT_HDR_SIZE -
-				     L2CAP_SDULEN_SIZE - L2CAP_FCS_SIZE);
+			hlen = L2CAP_EXT_HDR_SIZE + L2CAP_SDULEN_SIZE + L2CAP_FCS_SIZE;
+
+			if (chan->conn->mtu < hlen) {
+				BT_WARN("mtu is too small (mtu %d < len %d)",
+					chan->conn->mtu, hlen);
+				size = 0;
+			} else {
+				size = min_t(u16, le16_to_cpu(rfc.max_pdu_size),
+					     chan->conn->mtu - hlen);
+			}
+
 			rfc.max_pdu_size = cpu_to_le16(size);
 			chan->remote_mps = size;