[PATCH net v3] usbnet: fix EVENT_RX_PAUSED race in usbnet_skb_return()

[PATCH net v3] usbnet: fix EVENT_RX_PAUSED race in usbnet_skb_return()

[Ömer Mete Kaya]

On SMP, usbnet_skb_return() can pass the test_bit(EVENT_RX_PAUSED) check
on CPU0 while usbnet_resume_rx() clears the flag and drains rxq_pause on
CPU1, leaving the skb stranded until the next MTU change or device stop.

Protect the flag check and enqueue with a double-check pattern under
rxq_pause.lock on the slow path, keeping the normal RX fast path
lock-free. Wrap clear_bit() in usbnet_resume_rx() under the same lock.

Fixes: 43daa96b166c ("usbnet: Stop RX Q on MTU change")
Signed-off-by: Ömer Mete Kaya <omermetekaya0@gmail.com>

---
v3: Use spin_lock() inside existing local_bh_disable() section in
    usbnet_resume_rx() instead of spin_lock_bh(); move local_bh_enable()
    before queue_work() to restore BH context correctly.

 drivers/net/usb/usbnet.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)

diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c
index 84f97f448b2d..d138d353fd98 100644
--- a/drivers/net/usb/usbnet.c
+++ b/drivers/net/usb/usbnet.c
@@ -329,9 +329,14 @@ void usbnet_skb_return(struct usbnet *dev, struct sk_buff *skb)
 	unsigned long flags;
 	int	status;

-	if (test_bit(EVENT_RX_PAUSED, &dev->flags)) {
-		skb_queue_tail(&dev->rxq_pause, skb);
-		return;
+	if (unlikely(test_bit(EVENT_RX_PAUSED, &dev->flags))) {
+		spin_lock_bh(&dev->rxq_pause.lock);
+		if (test_bit(EVENT_RX_PAUSED, &dev->flags)) {
+			__skb_queue_tail(&dev->rxq_pause, skb);
+			spin_unlock_bh(&dev->rxq_pause.lock);
+			return;
+		}
+		spin_unlock_bh(&dev->rxq_pause.lock);
 	}

 	/* only update if unset to allow minidriver rx_fixup override */
@@ -702,15 +707,16 @@ void usbnet_resume_rx(struct usbnet *dev)
 	int num = 0;

 	local_bh_disable();
+	spin_lock(&dev->rxq_pause.lock);
 	clear_bit(EVENT_RX_PAUSED, &dev->flags);
-
+	spin_unlock(&dev->rxq_pause.lock);
 	while ((skb = skb_dequeue(&dev->rxq_pause)) != NULL) {
 		usbnet_skb_return(dev, skb);
 		num++;
 	}

-	queue_work(system_bh_wq, &dev->bh_work);
 	local_bh_enable();
+	queue_work(system_bh_wq, &dev->bh_work);

 	netif_dbg(dev, rx_status, dev->net,
 		  "paused rx queue disabled, %d skbs requeued\n", num);
--
2.56.0

Re: [PATCH net v3] usbnet: fix EVENT_RX_PAUSED race in usbnet_skb_return()

[netdev-bot+sinfo]

Hi!

This is an automated message. This series looks like a fix, but its
commit messages seem to be missing some information:

 - How the issue was discovered, e.g. hit in production, hit during
   development, syzbot report, manual code inspection, LLM or static
   analysis tool scan.

 - Whether the issue was actually triggered, or is only theoretical
   (e.g. found by code inspection). If it was triggered please include
   the symptoms, like the stack trace or error messages.

Please do not repost the series just to address the above. Instead,
reply to this email with the missing information, so that reviewers
can take it into account. If the series needs another revision for
other reasons, please include the information in the commit messages
then.

The evaluation is done by an LLM so it may be wrong, if you think
that is the case please reply and explain.

Re: [PATCH net v3] usbnet: fix EVENT_RX_PAUSED race in usbnet_skb_return()

[Jakub Kicinski]

On Wed, 07 Oct 2026 22:09:08 +0000 netdev-bot+sinfo@kernel.org wrote:
> This is an automated message. This series looks like a fix, but its
> commit messages seem to be missing some information:
> 
>  - How the issue was discovered, e.g. hit in production, hit during
>    development, syzbot report, manual code inspection, LLM or static
>    analysis tool scan.
> 
>  - Whether the issue was actually triggered, or is only theoretical
>    (e.g. found by code inspection). If it was triggered please include
>    the symptoms, like the stack trace or error messages.

Hi, please answer these questions, and try to include the answers in
the commit msg of future submissions.

Re: [PATCH net v3] usbnet: fix EVENT_RX_PAUSED race in usbnet_skb_return()

[Ömer Mete Kaya]

On 10/8/26 01:09, netdev-bot+sinfo@kernel.org wrote:
> Hi!
> 
> This is an automated message. This series looks like a fix, but its
> commit messages seem to be missing some information:
> 
>  - How the issue was discovered, e.g. hit in production, hit during
>    development, syzbot report, manual code inspection, LLM or static
>    analysis tool scan.

Sashiko AI suggestion.

>  - Whether the issue was actually triggered, or is only theoretical
>    (e.g. found by code inspection). If it was triggered please include
>    the symptoms, like the stack trace or error messages.

Theoretical.

> Please do not repost the series just to address the above. Instead,
> reply to this email with the missing information, so that reviewers
> can take it into account. If the series needs another revision for
> other reasons, please include the information in the commit messages
> then.
> 
> The evaluation is done by an LLM so it may be wrong, if you think
> that is the case please reply and explain.

Re: [PATCH net v3] usbnet: fix EVENT_RX_PAUSED race in usbnet_skb_return()

[Ömer Mete Kaya]

On 10/8/26 18:40, Jakub Kicinski wrote:
> Hi, please answer these questions, and try to include the answers in
> the commit msg of future submissions.

Hi, I couldnt add reported-by tag when it comes from sashiko,
should I include the answers under the "---" for future submissions?

Thanks.

Re: [PATCH net v3] usbnet: fix EVENT_RX_PAUSED race in usbnet_skb_return()

[Jakub Kicinski]

On Thu, 8 Oct 2026 18:47:07 +0300 Ömer Mete Kaya wrote:
> On 10/8/26 18:40, Jakub Kicinski wrote:
> > Hi, please answer these questions, and try to include the answers in
> > the commit msg of future submissions.  
> 
> Hi, I couldnt add reported-by tag when it comes from sashiko,
> should I include the answers under the "---" for future submissions?

In the main body is better, this is very useful info for risk / benefit
analysis for backporters.

Single sentence like this seems suitable here:

Discovered by Sashiko, not triggered, not tested on real HW.