[PATCH AUTOSEL 6.19-5.10] net: usb: r8152: fix transmit queue timeout

[Sasha Levin <sashal@kernel.org>]

From: Mingj Ye <insyelu@gmail.com>

[ Upstream commit 833dcd75d54f0bf5aa0a0781ff57456b421fbb40 ]

When the TX queue length reaches the threshold, the netdev watchdog
immediately detects a TX queue timeout.

This patch updates the trans_start timestamp of the transmit queue
on every asynchronous USB URB submission along the transmit path,
ensuring that the network watchdog accurately reflects ongoing
transmission activity.

Signed-off-by: Mingj Ye <insyelu@gmail.com>
Reviewed-by: Hayes Wang <hayeswang@realtek.com>
Link: https://patch.msgid.link/20260120015949.84996-1-insyelu@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---

LLM Generated explanations, may be completely bogus:

## Analysis of commit: "net: usb: r8152: fix transmit queue timeout"

### 1. COMMIT MESSAGE ANALYSIS

The subject explicitly says **"fix transmit queue timeout"** — this is a
clear bug fix. The commit message explains:
- **Bug**: When the TX queue length reaches a threshold, the netdev
  watchdog immediately detects a TX queue timeout (spurious timeout).
- **Root cause**: The `trans_start` timestamp of the transmit queue is
  not being updated when USB URBs are successfully submitted
  asynchronously.
- **Fix**: Call `netif_trans_update()` after each successful
  `usb_submit_urb()` to keep the watchdog informed that transmission
  activity is ongoing.

The commit is reviewed by the Realtek maintainer (Hayes Wang) and merged
by the network maintainer (Jakub Kicinski). Strong trust indicators.

### 2. CODE CHANGE ANALYSIS

The diff is **+2 lines** — extremely small and surgical:

```c
ret = usb_submit_urb(agg->urb, GFP_ATOMIC);
if (ret < 0)
    usb_autopm_put_interface_async(tp->intf);
+else
+    netif_trans_update(tp->netdev);
```

This adds an `else` branch after the `usb_submit_urb()` call. On
successful URB submission (`ret >= 0`), it calls
`netif_trans_update(tp->netdev)`, which updates the `trans_start`
timestamp of the network device's transmit queue.

**What `netif_trans_update()` does**: It's a standard kernel helper that
sets `txq->trans_start = jiffies`, telling the network watchdog "we are
actively transmitting." Without this update, the watchdog timer thinks
the queue has been idle since the last update and fires a spurious TX
timeout.

### 3. BUG SEVERITY

A **TX queue timeout** in a network driver is a significant user-visible
bug:
- The watchdog fires, potentially calling the driver's `ndo_tx_timeout`
  handler
- This can cause the network interface to reset or go down
- Results in **network connectivity loss** or **severe performance
  degradation**
- The r8152 driver is for **Realtek USB Ethernet adapters**, which are
  extremely common (used in USB-to-Ethernet dongles, docking stations,
  etc.)

### 4. SCOPE AND RISK ASSESSMENT

- **Lines changed**: 2 (adding an else branch with one function call)
- **Files touched**: 1 (`drivers/net/usb/r8152.c`)
- **Risk**: Extremely low. `netif_trans_update()` is a well-known,
  lightweight inline function. It simply updates a jiffies timestamp. It
  cannot cause crashes, data corruption, or any side effects.
- **Regression potential**: Near zero. This only affects the TX watchdog
  timer behavior and only updates the timestamp on *successful* URB
  submission.

### 5. USER IMPACT

- **HIGH impact** — r8152 is one of the most widely used USB Ethernet
  drivers. Realtek RTL8152/RTL8153 chips are in millions of USB Ethernet
  adapters.
- Users experience **spurious TX timeouts** that disrupt network
  connectivity.
- Affects any user with a Realtek USB Ethernet adapter under sufficient
  TX load.

### 6. STABLE KERNEL CRITERIA

| Criterion | Met? |
|-----------|------|
| Obviously correct and tested | YES — reviewed by Realtek maintainer,
simple logic |
| Fixes a real bug | YES — spurious TX queue timeouts |
| Important issue | YES — network connectivity loss |
| Small and contained | YES — 2 lines in one file |
| No new features | YES — just a watchdog timestamp update |
| No new APIs | YES |

### 7. DEPENDENCY CHECK

`netif_trans_update()` has been available in the kernel for a very long
time (since 4.7+). The r8152 driver and the `r8152_tx_agg_fill()`
function have been in stable trees for many kernel versions. This patch
applies cleanly with no dependencies on other commits.

### 8. CONCLUSION

This is a textbook stable backport candidate:
- Fixes a real, user-visible bug (spurious TX timeouts causing network
  disruption)
- Affects a widely-used USB Ethernet driver (Realtek r8152)
- Extremely small and surgical (2-line change)
- Zero regression risk (just updates a jiffies timestamp)
- Reviewed by the hardware vendor's maintainer
- No dependencies on other patches

**YES**

 drivers/net/usb/r8152.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c
index 2f3baa5f6e9c9..6b107cf5f37bd 100644
--- a/drivers/net/usb/r8152.c
+++ b/drivers/net/usb/r8152.c
@@ -2449,6 +2449,8 @@ static int r8152_tx_agg_fill(struct r8152 *tp, struct tx_agg *agg)
 	ret = usb_submit_urb(agg->urb, GFP_ATOMIC);
 	if (ret < 0)
 		usb_autopm_put_interface_async(tp->intf);
+	else
+		netif_trans_update(tp->netdev);
 
 out_tx_fill:
 	return ret;
-- 
2.51.0