[PATCH v10 0/3] net: can: isotp-fixes

[PATCH v10 0/3] net: can: isotp-fixes

[Oliver Hartkopp via B4 Relay]

As sashiko-bot was not able to check the second patch this bundle is
re-posted with b4 preparation.

Fixed the remarks reported by sashiko-bot of V9 patchset.

Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
---
Changes in v10:
- patch 3: add so->rx_lock support to sendmsg() and remove the former cmpxchg()
           concept to have so->tx.state's of relevant path's under lock
- formerly patch 4: removed as separate patch: integrated in patch 3
- Link to v9: https://patch.msgid.link/20260712-isotp-fixes-v9-0-137840d7f0f4@hartkopp.net

Changes in v9:
- patch 3: fix two reported issues and explain why issue 3&4 (-ECOMM sk_err
  handling) will not be changed
- Link to v8: https://patch.msgid.link/20260712-isotp-fixes-v8-0-c39220a4e442@hartkopp.net

Changes in v8:
- patch 3: properly terminate the former transmission before starting
- patch 4: formerly patch 5 as the former patch 4 is squashed into patch 3
- Link to v7: https://patch.msgid.link/20260711-isotp-fixes-v7-0-f03af24d8397@hartkopp.net

Changes in v7:
- patch 3: add helper so->tx_gen to monitor new ISOTP_SENDING attempts
- patch 4: prevent missasignment of sk_err report with new so->tx_gen
- Link to v6: https://patch.msgid.link/20260710-isotp-fixes-v6-0-a13c2539093d@hartkopp.net

Changes in v6:
- patch 2: check for rx state and tx state to be IDLE in isotp_bind() (sashiko-bot)
- patch 3: double check rx/tx states after all hrtimer_cancel() calls (sashiko-bot)
           move ISOTP_IDLE in err_event_drop for the same reason
- patch 4: use READ_ONCE() to prevent compiler load tearing as remarked by sashiko-bot
- Link to v5: https://patch.msgid.link/20260710-isotp-fixes-v5-0-15174d82d6f1@hartkopp.net

Changes in v5:
- patch 3: new: state change re-check after hrtimer_cancel()
- patch 4: new: handle state transistion with cmpxchg()
- Link to v4: https://patch.msgid.link/20260710-isotp-fixes-v4-0-2a4af437f61b@hartkopp.net

Changes in v4:
- patch 2: use a simpler method to reject a re-binding that can be implemented
  inside the lock: No waitqueue waiting just exit.
- Link to v3: https://patch.msgid.link/20260710-isotp-fixes-v3-0-08db68e27d0b@hartkopp.net

Changes in v3:
- patch 2: go back to initial idea without hard resetting the state machine and
  timers but let these come to a graceful end. Even a new bind() at netdev
  shutdown now leads to this graceful statemachine shutdown via the wait queue.
- Link to v2: https://patch.msgid.link/20260710-isotp-fixes-v2-0-bc57e26594b2@hartkopp.net

Changes in v2:
- patch 2: shutdown state machine to fix sashiko-bot complains
  Link: https://lore.kernel.org/linux-can/20260710094807.A50DD1F000E9@smtp.kernel.org/
- Link to v1: https://patch.msgid.link/20260710-isotp-fixes-v1-0-75a1d11d7df9@hartkopp.net

---
Oliver Hartkopp (3):
      can: isotp: use unconditional synchronize_rcu() in isotp_release()
      can: isotp: fix use-after-free race with concurrent NETDEV_UNREGISTER
      can: isotp: serialize TX state transitions under so->rx_lock

 net/can/isotp.c | 298 ++++++++++++++++++++++++++++++++++++++++++--------------
 1 file changed, 227 insertions(+), 71 deletions(-)
---
base-commit: dd3210c47e8d3ac6b4e9141fc68acc03b38c0ba3
change-id: 20260710-isotp-fixes-ead7480d67bf

Best regards,
--  
Oliver Hartkopp <socketcan@hartkopp.net>

[PATCH v10 1/3] can: isotp: use unconditional synchronize_rcu() in isotp_release()

[Oliver Hartkopp via B4 Relay]

From: Oliver Hartkopp <socketcan@hartkopp.net>

isotp_notify() unregisters the (RCU) CAN filters via can_rx_unregister()
and clears so->bound without waiting for a grace period. isotp_release()
uses so->bound to decide whether it needs to call synchronize_rcu()
before cancelling so->rxtimer, so when NETDEV_UNREGISTER runs first it
skips that synchronize_rcu() and can cancel the timer while an
in-flight isotp_rcv() is still executing and about to re-arm it via
isotp_send_fc(), leading to a use-after-free timer callback on the
freed socket.

sakisho-bot remarked a problem with rtnl_lock held in isotp_notify(),
therefore make isotp_release() always call synchronize_rcu() before
cancelling the timers, regardless of so->bound. This still closes the
original race (isotp_notify() clearing so->bound without waiting for
in-flight isotp_rcv() callers before isotp_release() cancels the RX
timer) without adding any RCU wait to the netdevice notifier path.

Fixes: 14a4696bc311 ("can: isotp: isotp_release(): omit unintended hrtimer restart on socket release")
Closes: https://lore.kernel.org/linux-can/20260707085210.6B6C01F000E9@smtp.kernel.org/
Reported-by: Nico Yip <zdi-disclosures@trendmicro.com> (ZDI-CAN-31764)
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
---
 net/can/isotp.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/net/can/isotp.c b/net/can/isotp.c
index c48b4a818297..d30937345bcd 100644
--- a/net/can/isotp.c
+++ b/net/can/isotp.c
@@ -1235,15 +1235,22 @@ static int isotp_release(struct socket *sock)
 
 				can_rx_unregister(net, dev, so->txid,
 						  SINGLE_MASK(so->txid),
 						  isotp_rcv_echo, sk);
 				dev_put(dev);
-				synchronize_rcu();
 			}
 		}
 	}
 
+	/* Always wait for a grace period before touching the timers below.
+	 * A concurrent NETDEV_UNREGISTER may have already unregistered our
+	 * filters and cleared so->bound in isotp_notify() without waiting
+	 * for in-flight isotp_rcv() callers to finish, so this call must not
+	 * be skipped just because so->bound is already 0 here.
+	 */
+	synchronize_rcu();
+
 	hrtimer_cancel(&so->txfrtimer);
 	hrtimer_cancel(&so->txtimer);
 	hrtimer_cancel(&so->rxtimer);
 
 	so->ifindex = 0;

-- 
2.53.0

[PATCH v10 2/3] can: isotp: fix use-after-free race with concurrent NETDEV_UNREGISTER

[Oliver Hartkopp via B4 Relay]

From: Oliver Hartkopp <socketcan@hartkopp.net>

isotp_release() looked up the bound network device via dev_get_by_index()
using the stored ifindex. During device unregistration the device is
unlisted from the ifindex hash before the NETDEV_UNREGISTER notifier
chain runs, so a concurrent isotp_release() could find no device, skip
can_rx_unregister() entirely, and still proceed to free the socket.
Since isotp_release() had already removed itself from the isotp
notifier list at that point, isotp_notify() would never get a chance to
clean up either, leaving a stale CAN filter that keeps pointing at the
freed socket.

Fix this the same way raw.c already does: hold a tracked reference to
the bound net_device in the socket (so->dev/so->dev_tracker) from
bind() onward instead of re-resolving it from the ifindex, and
serialize bind()/release() with rtnl_lock() so that so->dev is always
consistent with what the NETDEV_UNREGISTER notifier sees. so->dev
stays valid regardless of ifindex-hash unlisting, and is only ever
cleared by whichever of isotp_release()/isotp_notify() gets there
first, so the filter is always removed exactly once.

isotp_bind() now rejects a (re)bind with -EAGAIN while so->[tx|rx].state
isn't ISOTP_IDLE yet, so a timer left running by a prior
NETDEV_UNREGISTER can't act on a newly bound so->ifindex. Both checks
share the same lock_sock() section, so there is no window in which a
concurrent isotp_notify() clearing so->bound could be missed.

Fixes: e057dd3fc20f ("can: add ISO 15765-2:2016 transport protocol")
Reported-by: sashiko-bot@kernel.org
Link: https://lore.kernel.org/linux-can/20260707101420.47F261F000E9@smtp.kernel.org/
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
---
 net/can/isotp.c | 87 ++++++++++++++++++++++++++++++++++++++-------------------
 1 file changed, 59 insertions(+), 28 deletions(-)

diff --git a/net/can/isotp.c b/net/can/isotp.c
index d30937345bcd..44c044eb83e1 100644
--- a/net/can/isotp.c
+++ b/net/can/isotp.c
@@ -150,10 +150,12 @@ struct tpcon {
 
 struct isotp_sock {
 	struct sock sk;
 	int bound;
 	int ifindex;
+	struct net_device *dev;
+	netdevice_tracker dev_tracker;
 	canid_t txid;
 	canid_t rxid;
 	ktime_t tx_gap;
 	ktime_t lastrxcf_tstamp;
 	struct hrtimer rxtimer, txtimer, txfrtimer;
@@ -976,10 +978,18 @@ static int isotp_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
 		err = wait_event_interruptible(so->wait, so->tx.state == ISOTP_IDLE);
 		if (err)
 			goto err_event_drop;
 	}
 
+	/* so->bound is only checked once above - a wakeup may have
+	 * unbound/rebound the socket meanwhile, so re-validate it
+	 */
+	if (!so->bound) {
+		err = -EADDRNOTAVAIL;
+		goto err_out_drop;
+	}
+
 	/* PDU size > default => try max_pdu_size */
 	if (size > so->tx.buflen && so->tx.buflen < max_pdu_size) {
 		u8 *newbuf = kmalloc(max_pdu_size, GFP_KERNEL);
 
 		if (newbuf) {
@@ -1217,32 +1227,34 @@ static int isotp_release(struct socket *sock)
 		spin_lock(&isotp_notifier_lock);
 	}
 	list_del(&so->notifier);
 	spin_unlock(&isotp_notifier_lock);
 
+	rtnl_lock();
 	lock_sock(sk);
 
-	/* remove current filters & unregister */
-	if (so->bound) {
-		if (so->ifindex) {
-			struct net_device *dev;
-
-			dev = dev_get_by_index(net, so->ifindex);
-			if (dev) {
-				if (isotp_register_rxid(so))
-					can_rx_unregister(net, dev, so->rxid,
-							  SINGLE_MASK(so->rxid),
-							  isotp_rcv, sk);
-
-				can_rx_unregister(net, dev, so->txid,
-						  SINGLE_MASK(so->txid),
-						  isotp_rcv_echo, sk);
-				dev_put(dev);
-			}
-		}
+	/* remove current filters & unregister
+	 * tracked reference so->dev is taken at bind() time with rtnl_lock
+	 */
+	if (so->bound && so->dev) {
+		if (isotp_register_rxid(so))
+			can_rx_unregister(net, so->dev, so->rxid,
+					  SINGLE_MASK(so->rxid),
+					  isotp_rcv, sk);
+
+		can_rx_unregister(net, so->dev, so->txid,
+				  SINGLE_MASK(so->txid),
+				  isotp_rcv_echo, sk);
+		netdev_put(so->dev, &so->dev_tracker);
 	}
 
+	so->ifindex = 0;
+	so->bound = 0;
+	so->dev = NULL;
+
+	rtnl_unlock();
+
 	/* Always wait for a grace period before touching the timers below.
 	 * A concurrent NETDEV_UNREGISTER may have already unregistered our
 	 * filters and cleared so->bound in isotp_notify() without waiting
 	 * for in-flight isotp_rcv() callers to finish, so this call must not
 	 * be skipped just because so->bound is already 0 here.
@@ -1251,13 +1263,10 @@ static int isotp_release(struct socket *sock)
 
 	hrtimer_cancel(&so->txfrtimer);
 	hrtimer_cancel(&so->txtimer);
 	hrtimer_cancel(&so->rxtimer);
 
-	so->ifindex = 0;
-	so->bound = 0;
-
 	sock_orphan(sk);
 	sock->sk = NULL;
 
 	release_sock(sk);
 	sock_prot_inuse_add(net, sk->sk_prot, -1);
@@ -1308,17 +1317,29 @@ static int isotp_bind(struct socket *sock, struct sockaddr_unsized *uaddr, int l
 	}
 
 	if (!addr->can_ifindex)
 		return -ENODEV;
 
+	rtnl_lock();
 	lock_sock(sk);
 
 	if (so->bound) {
 		err = -EINVAL;
 		goto out;
 	}
 
+	/* A transmission or reception that outlived a previous binding
+	 * (unbound by NETDEV_UNREGISTER) may still be draining; the FC/echo
+	 * and RX watchdog timers bound how long this takes. Checked together
+	 * with so->bound in the same lock_sock() section above, so there is
+	 * no window in which a concurrent isotp_notify() could be missed.
+	 */
+	if (so->tx.state != ISOTP_IDLE || so->rx.state != ISOTP_IDLE) {
+		err = -EAGAIN;
+		goto out;
+	}
+
 	/* ensure different CAN IDs when the rx_id is to be registered */
 	if (isotp_register_rxid(so) && rx_id == tx_id) {
 		err = -EADDRNOTAVAIL;
 		goto out;
 	}
@@ -1327,18 +1348,16 @@ static int isotp_bind(struct socket *sock, struct sockaddr_unsized *uaddr, int l
 	if (!dev) {
 		err = -ENODEV;
 		goto out;
 	}
 	if (dev->type != ARPHRD_CAN) {
-		dev_put(dev);
 		err = -ENODEV;
-		goto out;
+		goto out_put_dev;
 	}
 	if (READ_ONCE(dev->mtu) < so->ll.mtu) {
-		dev_put(dev);
 		err = -EINVAL;
-		goto out;
+		goto out_put_dev;
 	}
 	if (!(dev->flags & IFF_UP))
 		notify_enetdown = 1;
 
 	ifindex = dev->ifindex;
@@ -1352,20 +1371,29 @@ static int isotp_bind(struct socket *sock, struct sockaddr_unsized *uaddr, int l
 
 	/* register for echo skb's */
 	can_rx_register(net, dev, tx_id, SINGLE_MASK(tx_id),
 			isotp_rcv_echo, sk, "isotpe", sk);
 
-	dev_put(dev);
-
 	/* switch to new settings */
 	so->ifindex = ifindex;
 	so->rxid = rx_id;
 	so->txid = tx_id;
 	so->bound = 1;
 
+	/* bind() ok -> hold a reference for so->dev so that isotp_release()
+	 * can safely reach the device later, even if a concurrent
+	 * NETDEV_UNREGISTER has already unlisted it by ifindex.
+	 */
+	so->dev = dev;
+	netdev_hold(so->dev, &so->dev_tracker, GFP_KERNEL);
+
+out_put_dev:
+	/* remove potential reference from dev_get_by_index() */
+	dev_put(dev);
 out:
 	release_sock(sk);
+	rtnl_unlock();
 
 	if (notify_enetdown) {
 		sk->sk_err = ENETDOWN;
 		if (!sock_flag(sk, SOCK_DEAD))
 			sk_error_report(sk);
@@ -1564,11 +1592,11 @@ static void isotp_notify(struct isotp_sock *so, unsigned long msg,
 	struct sock *sk = &so->sk;
 
 	if (!net_eq(dev_net(dev), sock_net(sk)))
 		return;
 
-	if (so->ifindex != dev->ifindex)
+	if (so->dev != dev)
 		return;
 
 	switch (msg) {
 	case NETDEV_UNREGISTER:
 		lock_sock(sk);
@@ -1580,14 +1608,16 @@ static void isotp_notify(struct isotp_sock *so, unsigned long msg,
 						  isotp_rcv, sk);
 
 			can_rx_unregister(dev_net(dev), dev, so->txid,
 					  SINGLE_MASK(so->txid),
 					  isotp_rcv_echo, sk);
+			netdev_put(so->dev, &so->dev_tracker);
 		}
 
 		so->ifindex = 0;
 		so->bound  = 0;
+		so->dev = NULL;
 		release_sock(sk);
 
 		sk->sk_err = ENODEV;
 		if (!sock_flag(sk, SOCK_DEAD))
 			sk_error_report(sk);
@@ -1643,10 +1673,11 @@ static int isotp_init(struct sock *sk)
 {
 	struct isotp_sock *so = isotp_sk(sk);
 
 	so->ifindex = 0;
 	so->bound = 0;
+	so->dev = NULL;
 
 	so->opt.flags = CAN_ISOTP_DEFAULT_FLAGS;
 	so->opt.ext_address = CAN_ISOTP_DEFAULT_EXT_ADDRESS;
 	so->opt.rx_ext_address = CAN_ISOTP_DEFAULT_EXT_ADDRESS;
 	so->opt.rxpad_content = CAN_ISOTP_DEFAULT_PAD_CONTENT;

-- 
2.53.0

[PATCH v10 3/3] can: isotp: serialize TX state transitions under so->rx_lock

[Oliver Hartkopp via B4 Relay]

From: Oliver Hartkopp <socketcan@hartkopp.net>

The TX state machine (so->tx.state) is driven from three contexts:
sendmsg() claiming and progressing a transfer, the RX path consuming
Flow Control/echo frames, and two hrtimers timing out a stalled
transfer. Mixing a lock-free cmpxchg() claim in sendmsg() with
hrtimer_cancel() calls made under so->rx_lock elsewhere left windows
where a frame or timer callback could act on a state that had already
moved on, corrupting an unrelated transfer.

so->rx_lock now covers the full lifecycle of a TX claim: sendmsg()
takes it to check so->tx.state is ISOTP_IDLE, switch it to
ISOTP_SENDING, bump so->tx_gen and drain the previous transfer's
timers - all as one critical section. isotp_rcv_fc()/isotp_rcv_cf()
already run under this lock via isotp_rcv(), and isotp_rcv_echo() now
takes it itself, so none of them can ever observe a transfer mid-claim.
This also means a transfer can no longer be handed to sendmsg()'s
cleanup paths (signal or send error) while another thread is
concurrently claiming or finishing it, so those paths can cancel
timers and reset the state unconditionally.

isotp_release() claims the socket the same way, so a racing sendmsg()
sees a consistent ISOTP_SHUTDOWN and skips arming its timer or sending.

Only the hrtimer callbacks stay outside so->rx_lock, since they run
under so->rx_lock's cancellation elsewhere and taking it themselves
would deadlock. so->tx_gen lets them recognize whether the transfer
they timed out is still the one currently active, so they don't
report an error against a transfer that has since completed or been
superseded.

Fixes: e057dd3fc20f ("can: add ISO 15765-2:2016 transport protocol")
Reported-by: sashiko-bot@kernel.org
Link: https://lore.kernel.org/linux-can/20260710142146.BDAE61F000E9@smtp.kernel.org/
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
---
 net/can/isotp.c | 202 ++++++++++++++++++++++++++++++++++++++++++++------------
 1 file changed, 160 insertions(+), 42 deletions(-)

diff --git a/net/can/isotp.c b/net/can/isotp.c
index 44c044eb83e1..54becaf6898f 100644
--- a/net/can/isotp.c
+++ b/net/can/isotp.c
@@ -156,18 +156,19 @@ struct isotp_sock {
 	netdevice_tracker dev_tracker;
 	canid_t txid;
 	canid_t rxid;
 	ktime_t tx_gap;
 	ktime_t lastrxcf_tstamp;
-	struct hrtimer rxtimer, txtimer, txfrtimer;
+	struct hrtimer rxtimer, txtimer, txfrtimer, echotimer;
 	struct can_isotp_options opt;
 	struct can_isotp_fc_options rxfc, txfc;
 	struct can_isotp_ll_options ll;
 	u32 frame_txtime;
 	u32 force_tx_stmin;
 	u32 force_rx_stmin;
 	u32 cfecho; /* consecutive frame echo tag */
+	u32 tx_gen; /* generation, bumped per new tx transfer */
 	struct tpcon rx, tx;
 	struct list_head notifier;
 	wait_queue_head_t wait;
 	spinlock_t rx_lock; /* protect single thread state machine */
 };
@@ -376,10 +377,19 @@ static int isotp_rcv_fc(struct isotp_sock *so, struct canfd_frame *cf, int ae)
 	    so->tx.state != ISOTP_WAIT_FIRST_FC)
 		return 0;
 
 	hrtimer_cancel(&so->txtimer);
 
+	/* isotp_tx_timeout() may have given up on this job while
+	 * hrtimer_cancel() above waited for it to finish; so->rx_lock
+	 * (held by our caller isotp_rcv()) rules out a concurrent claim,
+	 * so a plain recheck is enough here.
+	 */
+	if (so->tx.state != ISOTP_WAIT_FC &&
+	    so->tx.state != ISOTP_WAIT_FIRST_FC)
+		return 1;
+
 	if ((cf->len < ae + FC_CONTENT_SZ) ||
 	    ((so->opt.flags & ISOTP_CHECK_PADDING) &&
 	     check_pad(so, cf, ae + FC_CONTENT_SZ, so->opt.rxpad_content))) {
 		/* malformed PDU - report 'not a data message' */
 		sk->sk_err = EBADMSG;
@@ -422,11 +432,11 @@ static int isotp_rcv_fc(struct isotp_sock *so, struct canfd_frame *cf, int ae)
 	switch (cf->data[ae] & 0x0F) {
 	case ISOTP_FC_CTS:
 		so->tx.bs = 0;
 		so->tx.state = ISOTP_SENDING;
 		/* send CF frame and enable echo timeout handling */
-		hrtimer_start(&so->txtimer, ktime_set(ISOTP_ECHO_TIMEOUT, 0),
+		hrtimer_start(&so->echotimer, ktime_set(ISOTP_ECHO_TIMEOUT, 0),
 			      HRTIMER_MODE_REL_SOFT);
 		isotp_send_cframe(so);
 		break;
 
 	case ISOTP_FC_WT:
@@ -575,10 +585,18 @@ static int isotp_rcv_cf(struct sock *sk, struct canfd_frame *cf, int ae,
 		so->lastrxcf_tstamp = skb->tstamp;
 	}
 
 	hrtimer_cancel(&so->rxtimer);
 
+	/* isotp_rx_timer_handler() may have raced us for so->rx.state
+	 * while hrtimer_cancel() above waited for it to finish, already
+	 * reporting ETIMEDOUT and resetting the reception; don't process
+	 * this CF into a reassembly that has already been given up on.
+	 */
+	if (so->rx.state != ISOTP_WAIT_DATA)
+		return 1;
+
 	/* CFs are never longer than the FF */
 	if (cf->len > so->rx.ll_dl)
 		return 1;
 
 	/* CFs have usually the LL_DL length */
@@ -870,78 +888,124 @@ static void isotp_rcv_echo(struct sk_buff *skb, void *data)
 	struct sock *sk = (struct sock *)data;
 	struct isotp_sock *so = isotp_sk(sk);
 	struct canfd_frame *cf = (struct canfd_frame *)skb->data;
 
 	/* only handle my own local echo CF/SF skb's (no FF!) */
-	if (skb->sk != sk || so->cfecho != *(u32 *)cf->data)
+	if (skb->sk != sk)
 		return;
 
+	/* unlike isotp_rcv_fc()/isotp_rcv_cf(), not already under so->rx_lock
+	 * (no isotp_rcv() caller here), so take it ourselves
+	 */
+	spin_lock(&so->rx_lock);
+
+	/* so->cfecho may since belong to a new transfer; recheck under lock */
+	if (so->cfecho != *(u32 *)cf->data)
+		goto out_unlock;
+
 	/* cancel local echo timeout */
-	hrtimer_cancel(&so->txtimer);
+	hrtimer_cancel(&so->echotimer);
 
 	/* local echo skb with consecutive frame has been consumed */
 	so->cfecho = 0;
 
+	/* claiming a transfer also takes so->rx_lock, so a plain recheck
+	 * is enough: so->tx.state can't have flipped to ISOTP_SENDING for
+	 * a new claim while we're still in here
+	 */
+	if (so->tx.state != ISOTP_SENDING)
+		goto out_unlock;
+
 	if (so->tx.idx >= so->tx.len) {
 		/* we are done */
 		so->tx.state = ISOTP_IDLE;
 		wake_up_interruptible(&so->wait);
-		return;
+		goto out_unlock;
 	}
 
 	if (so->txfc.bs && so->tx.bs >= so->txfc.bs) {
 		/* stop and wait for FC with timeout */
 		so->tx.state = ISOTP_WAIT_FC;
 		hrtimer_start(&so->txtimer, ktime_set(ISOTP_FC_TIMEOUT, 0),
 			      HRTIMER_MODE_REL_SOFT);
-		return;
+		goto out_unlock;
 	}
 
 	/* no gap between data frames needed => use burst mode */
 	if (!so->tx_gap) {
 		/* enable echo timeout handling */
-		hrtimer_start(&so->txtimer, ktime_set(ISOTP_ECHO_TIMEOUT, 0),
+		hrtimer_start(&so->echotimer, ktime_set(ISOTP_ECHO_TIMEOUT, 0),
 			      HRTIMER_MODE_REL_SOFT);
 		isotp_send_cframe(so);
-		return;
+		goto out_unlock;
 	}
 
 	/* start timer to send next consecutive frame with correct delay */
 	hrtimer_start(&so->txfrtimer, so->tx_gap, HRTIMER_MODE_REL_SOFT);
+
+out_unlock:
+	spin_unlock(&so->rx_lock);
 }
 
-static enum hrtimer_restart isotp_tx_timer_handler(struct hrtimer *hrtimer)
+/* shared by so->txtimer's and so->echotimer's callbacks. Both timers get
+ * cancelled under so->rx_lock elsewhere, so this must stay lock-free to
+ * avoid deadlocking with that; uses so->tx_gen instead to avoid tainting
+ * a new transfer with an error from the one that just timed out.
+ */
+static enum hrtimer_restart isotp_tx_timeout(struct isotp_sock *so)
 {
-	struct isotp_sock *so = container_of(hrtimer, struct isotp_sock,
-					     txtimer);
 	struct sock *sk = &so->sk;
+	u32 gen = READ_ONCE(so->tx_gen);
+	u32 old_state = READ_ONCE(so->tx.state);
 
 	/* don't handle timeouts in IDLE or SHUTDOWN state */
-	if (so->tx.state == ISOTP_IDLE || so->tx.state == ISOTP_SHUTDOWN)
+	if (old_state == ISOTP_IDLE || old_state == ISOTP_SHUTDOWN)
+		return HRTIMER_NORESTART;
+
+	/* only claim the timeout if the state is still unchanged */
+	if (cmpxchg(&so->tx.state, old_state, ISOTP_IDLE) != old_state)
 		return HRTIMER_NORESTART;
 
 	/* we did not get any flow control or echo frame in time */
 
-	/* report 'communication error on send' */
-	sk->sk_err = ECOMM;
-	if (!sock_flag(sk, SOCK_DEAD))
-		sk_error_report(sk);
+	if (READ_ONCE(so->tx_gen) == gen) {
+		/* report 'communication error on send' */
+		sk->sk_err = ECOMM;
+		if (!sock_flag(sk, SOCK_DEAD))
+			sk_error_report(sk);
+	}
 
-	/* reset tx state */
-	so->tx.state = ISOTP_IDLE;
 	wake_up_interruptible(&so->wait);
 
 	return HRTIMER_NORESTART;
 }
 
+/* so->txtimer: fires when a Flow Control frame does not arrive in time */
+static enum hrtimer_restart isotp_tx_timer_handler(struct hrtimer *hrtimer)
+{
+	struct isotp_sock *so = container_of(hrtimer, struct isotp_sock,
+					     txtimer);
+
+	return isotp_tx_timeout(so);
+}
+
+/* so->echotimer: fires when a sent CF/SF's local echo does not arrive */
+static enum hrtimer_restart isotp_echo_timer_handler(struct hrtimer *hrtimer)
+{
+	struct isotp_sock *so = container_of(hrtimer, struct isotp_sock,
+					     echotimer);
+
+	return isotp_tx_timeout(so);
+}
+
 static enum hrtimer_restart isotp_txfr_timer_handler(struct hrtimer *hrtimer)
 {
 	struct isotp_sock *so = container_of(hrtimer, struct isotp_sock,
 					     txfrtimer);
 
 	/* start echo timeout handling and cover below protocol error */
-	hrtimer_start(&so->txtimer, ktime_set(ISOTP_ECHO_TIMEOUT, 0),
+	hrtimer_start(&so->echotimer, ktime_set(ISOTP_ECHO_TIMEOUT, 0),
 		      HRTIMER_MODE_REL_SOFT);
 
 	/* cfecho should be consumed by isotp_rcv_echo() here */
 	if (so->tx.state == ISOTP_SENDING && !so->cfecho)
 		isotp_send_cframe(so);
@@ -958,30 +1022,53 @@ static int isotp_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
 	struct net_device *dev;
 	struct canfd_frame *cf;
 	int ae = (so->opt.flags & CAN_ISOTP_EXTEND_ADDR) ? 1 : 0;
 	int wait_tx_done = (so->opt.flags & CAN_ISOTP_WAIT_TX_DONE) ? 1 : 0;
 	s64 hrtimer_sec = ISOTP_ECHO_TIMEOUT;
+	struct hrtimer *tx_hrt = &so->echotimer;
+	u32 new_state = ISOTP_SENDING;
 	int off;
 	int err;
 
 	if (!so->bound || so->tx.state == ISOTP_SHUTDOWN)
 		return -EADDRNOTAVAIL;
 
-	while (cmpxchg(&so->tx.state, ISOTP_IDLE, ISOTP_SENDING) != ISOTP_IDLE) {
+	/* claim the socket under so->rx_lock: this serializes the claim
+	 * with the RX path and with sendmsg()'s own error paths below, so
+	 * none of them can ever see a transfer mid-claim
+	 */
+	for (;;) {
+		spin_lock_bh(&so->rx_lock);
+		if (READ_ONCE(so->tx.state) == ISOTP_IDLE)
+			break;
+		spin_unlock_bh(&so->rx_lock);
+
 		/* we do not support multiple buffers - for now */
 		if (msg->msg_flags & MSG_DONTWAIT)
 			return -EAGAIN;
 
 		if (so->tx.state == ISOTP_SHUTDOWN)
 			return -EADDRNOTAVAIL;
 
 		/* wait for complete transmission of current pdu */
-		err = wait_event_interruptible(so->wait, so->tx.state == ISOTP_IDLE);
+		err = wait_event_interruptible(so->wait,
+					       so->tx.state == ISOTP_IDLE);
 		if (err)
-			goto err_event_drop;
+			return err;
 	}
 
+	/* new transfer: bump so->tx_gen and drain the old one's timers,
+	 * still under the so->rx_lock we just claimed the socket with
+	 */
+	WRITE_ONCE(so->tx.state, ISOTP_SENDING);
+	WRITE_ONCE(so->tx_gen, READ_ONCE(so->tx_gen) + 1);
+	hrtimer_cancel(&so->txtimer);
+	hrtimer_cancel(&so->echotimer);
+	hrtimer_cancel(&so->txfrtimer);
+	so->cfecho = 0;
+	spin_unlock_bh(&so->rx_lock);
+
 	/* so->bound is only checked once above - a wakeup may have
 	 * unbound/rebound the socket meanwhile, so re-validate it
 	 */
 	if (!so->bound) {
 		err = -EADDRNOTAVAIL;
@@ -1094,22 +1181,37 @@ static int isotp_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
 
 			/* set CF echo tag for isotp_rcv_echo() (CF-mode) */
 			so->cfecho = *(u32 *)cf->data;
 		} else {
 			/* standard flow control check */
-			so->tx.state = ISOTP_WAIT_FIRST_FC;
+			new_state = ISOTP_WAIT_FIRST_FC;
 
 			/* start timeout for FC */
 			hrtimer_sec = ISOTP_FC_TIMEOUT;
+			tx_hrt = &so->txtimer;
 
 			/* no CF echo tag for isotp_rcv_echo() (FF-mode) */
 			so->cfecho = 0;
 		}
 	}
 
-	hrtimer_start(&so->txtimer, ktime_set(hrtimer_sec, 0),
+	spin_lock_bh(&so->rx_lock);
+	if (so->tx.state == ISOTP_SHUTDOWN) {
+		/* isotp_release() has since taken over and already drained
+		 * our timers - don't send into a socket that's going away
+		 */
+		spin_unlock_bh(&so->rx_lock);
+		kfree_skb(skb);
+		dev_put(dev);
+		wake_up_interruptible(&so->wait);
+		return -EADDRNOTAVAIL;
+	}
+	/* WAIT_FIRST_FC for standard FF, else stays ISOTP_SENDING */
+	so->tx.state = new_state;
+	hrtimer_start(tx_hrt, ktime_set(hrtimer_sec, 0),
 		      HRTIMER_MODE_REL_SOFT);
+	spin_unlock_bh(&so->rx_lock);
 
 	/* send the first or only CAN frame */
 	cf->flags = so->ll.tx_flags;
 
 	skb->dev = dev;
@@ -1118,17 +1220,14 @@ static int isotp_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
 	dev_put(dev);
 	if (err) {
 		pr_notice_once("can-isotp: %s: can_send_ret %pe\n",
 			       __func__, ERR_PTR(err));
 
+		spin_lock_bh(&so->rx_lock);
 		/* no transmission -> no timeout monitoring */
-		hrtimer_cancel(&so->txtimer);
-
-		/* reset consecutive frame echo tag */
-		so->cfecho = 0;
-
-		goto err_out_drop;
+		hrtimer_cancel(tx_hrt);
+		goto err_out_drop_locked;
 	}
 
 	if (wait_tx_done) {
 		/* wait for complete transmission of current pdu */
 		err = wait_event_interruptible(so->wait, so->tx.state == ISOTP_IDLE);
@@ -1140,18 +1239,25 @@ static int isotp_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
 			return err;
 	}
 
 	return size;
 
+err_out_drop:
+	/* claimed but nothing sent yet - no timer to cancel */
+	spin_lock_bh(&so->rx_lock);
+	goto err_out_drop_locked;
 err_event_drop:
-	/* got signal: force tx state machine to be idle */
-	so->tx.state = ISOTP_IDLE;
+	/* interrupted waiting on our own transfer - drain its timers */
+	spin_lock_bh(&so->rx_lock);
 	hrtimer_cancel(&so->txfrtimer);
 	hrtimer_cancel(&so->txtimer);
-err_out_drop:
-	/* drop this PDU and unlock a potential wait queue */
+	hrtimer_cancel(&so->echotimer);
+err_out_drop_locked:
+	/* release the claim; so->rx_lock still held from above */
+	so->cfecho = 0;
 	so->tx.state = ISOTP_IDLE;
+	spin_unlock_bh(&so->rx_lock);
 	wake_up_interruptible(&so->wait);
 
 	return err;
 }
 
@@ -1209,17 +1315,24 @@ static int isotp_release(struct socket *sock)
 		return 0;
 
 	so = isotp_sk(sk);
 	net = sock_net(sk);
 
-	/* wait for complete transmission of current pdu */
-	while (wait_event_interruptible(so->wait, so->tx.state == ISOTP_IDLE) == 0 &&
-	       cmpxchg(&so->tx.state, ISOTP_IDLE, ISOTP_SHUTDOWN) != ISOTP_IDLE)
+	/* best-effort: wait for a running pdu to finish, but don't block on
+	 * it forever - give up after the first signal
+	 */
+	while (so->tx.state != ISOTP_IDLE &&
+	       wait_event_interruptible(so->wait, so->tx.state == ISOTP_IDLE) == 0)
 		;
 
-	/* force state machines to be idle also when a signal occurred */
+	/* claim the socket under so->rx_lock like sendmsg() does, so its
+	 * claim can't race the forced ISOTP_SHUTDOWN below; force it
+	 * unconditionally, even when a signal cut the wait above short
+	 */
+	spin_lock_bh(&so->rx_lock);
 	so->tx.state = ISOTP_SHUTDOWN;
+	spin_unlock_bh(&so->rx_lock);
 	so->rx.state = ISOTP_IDLE;
 
 	spin_lock(&isotp_notifier_lock);
 	while (isotp_busy_notifier == so) {
 		spin_unlock(&isotp_notifier_lock);
@@ -1261,10 +1374,11 @@ static int isotp_release(struct socket *sock)
 	 */
 	synchronize_rcu();
 
 	hrtimer_cancel(&so->txfrtimer);
 	hrtimer_cancel(&so->txtimer);
+	hrtimer_cancel(&so->echotimer);
 	hrtimer_cancel(&so->rxtimer);
 
 	sock_orphan(sk);
 	sock->sk = NULL;
 
@@ -1700,14 +1814,18 @@ static int isotp_init(struct sock *sk)
 	so->rx.buf = so->rx.sbuf;
 	so->tx.buf = so->tx.sbuf;
 	so->rx.buflen = ARRAY_SIZE(so->rx.sbuf);
 	so->tx.buflen = ARRAY_SIZE(so->tx.sbuf);
 
-	hrtimer_setup(&so->rxtimer, isotp_rx_timer_handler, CLOCK_MONOTONIC, HRTIMER_MODE_REL_SOFT);
-	hrtimer_setup(&so->txtimer, isotp_tx_timer_handler, CLOCK_MONOTONIC, HRTIMER_MODE_REL_SOFT);
-	hrtimer_setup(&so->txfrtimer, isotp_txfr_timer_handler, CLOCK_MONOTONIC,
-		      HRTIMER_MODE_REL_SOFT);
+	hrtimer_setup(&so->rxtimer, isotp_rx_timer_handler,
+		      CLOCK_MONOTONIC, HRTIMER_MODE_REL_SOFT);
+	hrtimer_setup(&so->txtimer, isotp_tx_timer_handler,
+		      CLOCK_MONOTONIC, HRTIMER_MODE_REL_SOFT);
+	hrtimer_setup(&so->echotimer, isotp_echo_timer_handler,
+		      CLOCK_MONOTONIC, HRTIMER_MODE_REL_SOFT);
+	hrtimer_setup(&so->txfrtimer, isotp_txfr_timer_handler,
+		      CLOCK_MONOTONIC, HRTIMER_MODE_REL_SOFT);
 
 	init_waitqueue_head(&so->wait);
 	spin_lock_init(&so->rx_lock);
 
 	spin_lock(&isotp_notifier_lock);

-- 
2.53.0

Re: [PATCH v10 3/3] can: isotp: serialize TX state transitions under so->rx_lock

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Lockless state transition in the timeout handler creates a race condition that causes timeout errors to be silently dropped or misattributed, leading sendmsg() to erroneously return success for timed-out transfers.
--

--- Patch [3]: [PATCH v10 3/3] can: isotp: serialize TX state transitions under so->rx_lock ---
commit e21ce85bab56b1b57ffd11ce2b23b760bfb72c21
Author: Oliver Hartkopp <socketcan@hartkopp.net>

can: isotp: serialize TX state transitions under so->rx_lock

This commit serializes CAN ISOTP TX state transitions and timer cancellations
under so->rx_lock to prevent concurrent modifications. It also changes the
timeout handlers to locklessly cancel transfers by checking and updating
so->tx_gen and so->tx.state.

Link: https://lore.kernel.org/linux-can/20260710142146.BDAE61F000E9@smtp.kernel.org/

> diff --git a/net/can/isotp.c b/net/can/isotp.c
> index 44c044eb83e12..54becaf6898f1 100644
> --- a/net/can/isotp.c
> +++ b/net/can/isotp.c
[ ... ]
> @@ -893,53 +927,83 @@ static void isotp_rcv_echo(struct sk_buff *skb, void *data)
[ ... ]
> +static enum hrtimer_restart isotp_tx_timeout(struct isotp_sock *so)
>  {
> -	struct isotp_sock *so = container_of(hrtimer, struct isotp_sock,
> -					     txtimer);
>  	struct sock *sk = &so->sk;
> +	u32 gen = READ_ONCE(so->tx_gen);
> +	u32 old_state = READ_ONCE(so->tx.state);
>  
>  	/* don't handle timeouts in IDLE or SHUTDOWN state */
> -	if (so->tx.state == ISOTP_IDLE || so->tx.state == ISOTP_SHUTDOWN)
> +	if (old_state == ISOTP_IDLE || old_state == ISOTP_SHUTDOWN)
> +		return HRTIMER_NORESTART;
> +
> +	/* only claim the timeout if the state is still unchanged */
> +	if (cmpxchg(&so->tx.state, old_state, ISOTP_IDLE) != old_state)
>  		return HRTIMER_NORESTART;
>  
>  	/* we did not get any flow control or echo frame in time */
>  
> -	/* report 'communication error on send' */
> -	sk->sk_err = ECOMM;
> -	if (!sock_flag(sk, SOCK_DEAD))
> -		sk_error_report(sk);
> +	if (READ_ONCE(so->tx_gen) == gen) {
> +		/* report 'communication error on send' */
> +		sk->sk_err = ECOMM;
> +		if (!sock_flag(sk, SOCK_DEAD))
> +			sk_error_report(sk);
> +	}

[Severity: High]
Does this lockless state transition create a race window where timeout
errors are lost?

By setting so->tx.state to ISOTP_IDLE before setting sk->sk_err = ECOMM,
a thread waiting in isotp_sendmsg() for wait_tx_done might wake up, observe
ISOTP_IDLE, and read sock_error() before the error is set, falsely returning
success.

Additionally, could another concurrent sendmsg() see the IDLE state,
claim the socket, and increment tx_gen before the error is set?

If this happens, the check if (READ_ONCE(so->tx_gen) == gen) will fail,
and the error will never be reported to the original waiting thread.

[ ... ]
> @@ -960,13 +1024,24 @@ static int isotp_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
[ ... ]
> +	/* new transfer: bump so->tx_gen and drain the old one's timers,
> +	 * still under the so->rx_lock we just claimed the socket with
> +	 */
> +	WRITE_ONCE(so->tx.state, ISOTP_SENDING);
> +	WRITE_ONCE(so->tx_gen, READ_ONCE(so->tx_gen) + 1);
> +	hrtimer_cancel(&so->txtimer);
> +	hrtimer_cancel(&so->echotimer);

[Severity: High]
When this tx_gen increment occurs concurrently with isotp_tx_timeout()
after the timeout has already transitioned the state to ISOTP_IDLE, doesn't it
cause the timeout handler to suppress the error for the previous transfer?

[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260712-isotp-fixes-v10-0-793a1b1ce17f@hartkopp.net?part=3

Re: [PATCH v10 1/3] can: isotp: use unconditional synchronize_rcu() in isotp_release()

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:

Pre-existing issues:
- [High] CAN filter leak in isotp_release() during concurrent netdevice unregistration leading to Use-After-Free.
- [High] Missing synchronization and state cleanup in isotp_sendmsg error path causes data corruption when racing with CAN echo frames.
--

--- Patch [1]: [PATCH v10 1/3] can: isotp: use unconditional synchronize_rcu() in isotp_release() ---
commit fa3db7e87e5eabd37c1bb7bc38aae86169f62182
Author: Oliver Hartkopp <socketcan@hartkopp.net>

can: isotp: use unconditional synchronize_rcu() in isotp_release()

This commit adds an unconditional synchronize_rcu() in isotp_release()
before cancelling timers. It addresses a race condition where a
concurrent NETDEV_UNREGISTER could cause a use-after-free in the timer
callback on the freed socket.

Closes: https://lore.kernel.org/linux-can/20260707085210.6B6C01F000E9@smtp.kernel.org/

> diff --git a/net/can/isotp.c b/net/can/isotp.c
> index c48b4a818297e..d30937345bcdc 100644
> --- a/net/can/isotp.c
> +++ b/net/can/isotp.c
> @@ -1237,11 +1237,18 @@ static int isotp_release(struct socket *sock)
>  						  SINGLE_MASK(so->txid),
>  						  isotp_rcv_echo, sk);
>  				dev_put(dev);

[Severity: High]
This is a pre-existing issue, but is there a potential CAN filter leak here
that could lead to a use-after-free?

If a user closes the ISOTP socket exactly when the bound netdevice is being
unregistered, unlist_netdevice(dev) is called before the NETDEV_UNREGISTER
notifiers.

If isotp_release() runs in this window, it removes the socket from
isotp_notifier_list and calls dev_get_by_index(). Since the device is
unlisted, dev_get_by_index() returns NULL, causing isotp_release() to skip
can_rx_unregister().

Because the filter is left active, any matching CAN frames subsequently
processed by the unregistering device might trigger isotp_rcv() with the
freed socket pointer. Does the newly added synchronize_rcu() below prevent
this if the filter is never actually unregistered?

> -				synchronize_rcu();
>  			}
>  		}
>  	}
>  
> +	/* Always wait for a grace period before touching the timers below.
> +	 * A concurrent NETDEV_UNREGISTER may have already unregistered our
> +	 * filters and cleared so->bound in isotp_notify() without waiting
> +	 * for in-flight isotp_rcv() callers to finish, so this call must not
> +	 * be skipped just because so->bound is already 0 here.
> +	 */
> +	synchronize_rcu();
> +
>  	hrtimer_cancel(&so->txfrtimer);
>  	hrtimer_cancel(&so->txtimer);
>  	hrtimer_cancel(&so->rxtimer);

[Severity: High]
This is a pre-existing issue, but does the error path in isotp_sendmsg()
properly clean up the echo state when interrupted by a signal?

If a user initiates a blocking transfer with wait_tx_done enabled, and it
is interrupted by a signal, the err_event_drop path forcefully sets
so->tx.state = ISOTP_IDLE without clearing so->cfecho or waiting for the
in-flight CAN hardware echo:

net/can/isotp.c:isotp_sendmsg() {
[ ... ]
err_event_drop:
        /* got signal: force tx state machine to be idle */
        so->tx.state = ISOTP_IDLE;
        hrtimer_cancel(&so->txfrtimer);
        hrtimer_cancel(&so->txtimer);
[ ... ]
}

When the background echo frame arrives, isotp_rcv_echo() matches the stale
so->cfecho and unconditionally continues transmission by reading from
so->tx.buf and incrementing so->tx.idx:

net/can/isotp.c:isotp_rcv_echo() {
[ ... ]
        if (skb->sk != sk || so->cfecho != *(u32 *)cf->data)
                return;
[ ... ]
        if (!so->tx_gap) {
                /* enable echo timeout handling */
                hrtimer_start(&so->txtimer, ktime_set(ISOTP_ECHO_TIMEOUT, 0),
                              HRTIMER_MODE_REL_SOFT);
                isotp_send_cframe(so);
                return;
        }
[ ... ]
}

Concurrently, the user can call sendmsg() again since the state is IDLE,
overwriting the buffer and lengths without locks. Could this time-of-check
to time-of-use race cause the background softirq to read corrupted indices
and result in out-of-bounds memory accesses?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260712-isotp-fixes-v10-0-793a1b1ce17f@hartkopp.net?part=1

Re: [PATCH v10 3/3] can: isotp: serialize TX state transitions under so->rx_lock

[Oliver Hartkopp]

On 12.07.26 20:11, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [High] Lockless state transition in the timeout handler creates a race condition that causes timeout errors to be silently dropped or misattributed, leading sendmsg() to erroneously return success for timed-out transfers.

The patch set fixes potential UAF and stalls.

The error return codes especially for (uncommon) multi-threaded users 
have been there for years and might be fixed in later patches.

So no reason to delay this patch set.

Best regards,
Oliver

> --
> 
> --- Patch [3]: [PATCH v10 3/3] can: isotp: serialize TX state transitions under so->rx_lock ---
> commit e21ce85bab56b1b57ffd11ce2b23b760bfb72c21
> Author: Oliver Hartkopp <socketcan@hartkopp.net>
> 
> can: isotp: serialize TX state transitions under so->rx_lock
> 
> This commit serializes CAN ISOTP TX state transitions and timer cancellations
> under so->rx_lock to prevent concurrent modifications. It also changes the
> timeout handlers to locklessly cancel transfers by checking and updating
> so->tx_gen and so->tx.state.
> 
> Link: https://lore.kernel.org/linux-can/20260710142146.BDAE61F000E9@smtp.kernel.org/
> 
>> diff --git a/net/can/isotp.c b/net/can/isotp.c
>> index 44c044eb83e12..54becaf6898f1 100644
>> --- a/net/can/isotp.c
>> +++ b/net/can/isotp.c
> [ ... ]
>> @@ -893,53 +927,83 @@ static void isotp_rcv_echo(struct sk_buff *skb, void *data)
> [ ... ]
>> +static enum hrtimer_restart isotp_tx_timeout(struct isotp_sock *so)
>>   {
>> -	struct isotp_sock *so = container_of(hrtimer, struct isotp_sock,
>> -					     txtimer);
>>   	struct sock *sk = &so->sk;
>> +	u32 gen = READ_ONCE(so->tx_gen);
>> +	u32 old_state = READ_ONCE(so->tx.state);
>>   
>>   	/* don't handle timeouts in IDLE or SHUTDOWN state */
>> -	if (so->tx.state == ISOTP_IDLE || so->tx.state == ISOTP_SHUTDOWN)
>> +	if (old_state == ISOTP_IDLE || old_state == ISOTP_SHUTDOWN)
>> +		return HRTIMER_NORESTART;
>> +
>> +	/* only claim the timeout if the state is still unchanged */
>> +	if (cmpxchg(&so->tx.state, old_state, ISOTP_IDLE) != old_state)
>>   		return HRTIMER_NORESTART;
>>   
>>   	/* we did not get any flow control or echo frame in time */
>>   
>> -	/* report 'communication error on send' */
>> -	sk->sk_err = ECOMM;
>> -	if (!sock_flag(sk, SOCK_DEAD))
>> -		sk_error_report(sk);
>> +	if (READ_ONCE(so->tx_gen) == gen) {
>> +		/* report 'communication error on send' */
>> +		sk->sk_err = ECOMM;
>> +		if (!sock_flag(sk, SOCK_DEAD))
>> +			sk_error_report(sk);
>> +	}
> 
> [Severity: High]
> Does this lockless state transition create a race window where timeout
> errors are lost?
> 
> By setting so->tx.state to ISOTP_IDLE before setting sk->sk_err = ECOMM,
> a thread waiting in isotp_sendmsg() for wait_tx_done might wake up, observe
> ISOTP_IDLE, and read sock_error() before the error is set, falsely returning
> success.
> 
> Additionally, could another concurrent sendmsg() see the IDLE state,
> claim the socket, and increment tx_gen before the error is set?
> 
> If this happens, the check if (READ_ONCE(so->tx_gen) == gen) will fail,
> and the error will never be reported to the original waiting thread.
> 
> [ ... ]
>> @@ -960,13 +1024,24 @@ static int isotp_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
> [ ... ]
>> +	/* new transfer: bump so->tx_gen and drain the old one's timers,
>> +	 * still under the so->rx_lock we just claimed the socket with
>> +	 */
>> +	WRITE_ONCE(so->tx.state, ISOTP_SENDING);
>> +	WRITE_ONCE(so->tx_gen, READ_ONCE(so->tx_gen) + 1);
>> +	hrtimer_cancel(&so->txtimer);
>> +	hrtimer_cancel(&so->echotimer);
> 
> [Severity: High]
> When this tx_gen increment occurs concurrently with isotp_tx_timeout()
> after the timeout has already transitioned the state to ISOTP_IDLE, doesn't it
> cause the timeout handler to suppress the error for the previous transfer?
> 
> [ ... ]
>