Thread (10 messages) 10 messages, 3 authors, 2025-07-23
STALE367d

[PATCH] net: usbnet: Avoid potential RCU stall on LINK_CHANGE event

From: John Ernberg <hidden>
Date: 2025-07-10 09:05:52
Also in: linux-usb, lkml, stable
Subsystem: networking drivers, the rest, usb "usbnet" driver framework, usb networking drivers · Maintainers: Andrew Lunn, "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Linus Torvalds, Oliver Neukum

Having a Gemalto Cinterion PLS83-W modem attached to USB and activating the
cellular data link would sometimes yield the following RCU stall, leading
to a system freeze:

    rcu: INFO: rcu_sched detected expedited stalls on CPUs/tasks: { 0-.... } 33108 jiffies s: 201 root: 0x1/.
    rcu: blocking rcu_node structures (internal RCU debug):
    Sending NMI from CPU 1 to CPUs 0:
    NMI backtrace for cpu 0

    Call trace:
     arch_local_irq_enable+0x4/0x8
     local_bh_enable+0x18/0x20
     __netdev_alloc_skb+0x18c/0x1cc
     rx_submit+0x68/0x1f8 [usbnet]
     rx_alloc_submit+0x4c/0x74 [usbnet]
     usbnet_bh+0x1d8/0x218 [usbnet]
     usbnet_bh_tasklet+0x10/0x18 [usbnet]
     tasklet_action_common+0xa8/0x110
     tasklet_action+0x2c/0x34
     handle_softirqs+0x2cc/0x3a0
     __do_softirq+0x10/0x18
     ____do_softirq+0xc/0x14
     call_on_irq_stack+0x24/0x34
     do_softirq_own_stack+0x18/0x20
     __irq_exit_rcu+0xa8/0xb8
     irq_exit_rcu+0xc/0x30
     el1_interrupt+0x34/0x48
     el1h_64_irq_handler+0x14/0x1c
     el1h_64_irq+0x68/0x6c
     _raw_spin_unlock_irqrestore+0x38/0x48
     xhci_urb_dequeue+0x1ac/0x45c [xhci_hcd]
     unlink1+0xd4/0xdc [usbcore]
     usb_hcd_unlink_urb+0x70/0xb0 [usbcore]
     usb_unlink_urb+0x24/0x44 [usbcore]
     unlink_urbs.constprop.0.isra.0+0x64/0xa8 [usbnet]
     __handle_link_change+0x34/0x70 [usbnet]
     usbnet_deferred_kevent+0x1c0/0x320 [usbnet]
     process_scheduled_works+0x2d0/0x48c
     worker_thread+0x150/0x1dc
     kthread+0xd8/0xe8
     ret_from_fork+0x10/0x20

It turns out that during the link activation a LINK_CHANGE event is emitted
which causes the active RX URBs to be unlinked, while that is happening
rx_submit() may begin pushing new URBs to the queue being emptied.
Causing the unlink queue to never empty.

Use the same approach as commit 43daa96b166c ("usbnet: Stop RX Q on MTU
change") and pause the RX queue while unlinking the URBs on LINK_CHANGE
as well.

Fixes: 4b49f58fff00 ("usbnet: handle link change")
Cc: stable@vger.kernel.org
Signed-off-by: John Ernberg <redacted>

---

Tested on 6.12.20 and forward ported.
---
 drivers/net/usb/usbnet.c | 2 ++
 1 file changed, 2 insertions(+)
diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c
index c04e715a4c2a..156f0e85a135 100644
--- a/drivers/net/usb/usbnet.c
+++ b/drivers/net/usb/usbnet.c
@@ -1115,7 +1115,9 @@ static void __handle_link_change(struct usbnet *dev)
 
 	if (!netif_carrier_ok(dev->net)) {
 		/* kill URBs for reading packets to save bus bandwidth */
+		usbnet_pause_rx(dev);
 		unlink_urbs(dev, &dev->rxq);
+		usbnet_resume_rx(dev);
 
 		/*
 		 * tx_timeout will unlink URBs for sending packets and
-- 
2.49.0
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help