[PATCH net] net: bcmgenet: stop Tx NAPI before disabling the queues

bcmgenet_netif_stop() disables the Tx queues first and stops Tx NAPI
several steps later. A completion in flight calls netif_tx_wake_queue() in
between, so a transmit can reach the rings after bcmgenet_fini_dma() has
freed them.

Close is safe because dev_deactivate_many() stops the qdisc first.
bcmgenet_suspend() does not, so stop Tx NAPI before the queues.

KASAN on a Raspberry Pi CM4, driven from an MTU change because suspend
freezes user space before the callback runs:

  BUG: KASAN: use-after-free in bcmgenet_xmit+0x17f8/0x2258
  Write of size 8 at addr ffffff8055844a68 by task ksoftirqd/0/14
   bcmgenet_xmit+0x17f8/0x2258
   dev_hard_start_xmit+0x13c/0x588
   sch_direct_xmit+0x108/0x340
   __dev_queue_xmit+0x1190/0x3848

Fixes: 254f3239dd07 ("net: bcmgenet: revise suspend/resume")
Signed-off-by: Nicolai Buchwitz <redacted>
---
Found this, while I was preparing my jumbo frame series for genet.
The MTU change in my test comes from that series and is not in tree
yet.

 drivers/net/ethernet/broadcom/genet/bcmgenet.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/broadcom/genet/bcmgenet.c b/drivers/net/ethernet/broadcom/genet/bcmgenet.c
index b916080f4ff1..0789b60e664a 100644
--- a/drivers/net/ethernet/broadcom/genet/bcmgenet.c
+++ b/drivers/net/ethernet/broadcom/genet/bcmgenet.c

@@ -3441,6 +3441,8 @@ static void bcmgenet_netif_stop(struct net_device *dev, bool stop_phy)
 {
 	struct bcmgenet_priv *priv = netdev_priv(dev);
 
+	/* Stop completion polling before it can wake a stopped queue */
+	bcmgenet_disable_tx_napi(priv);
 	netif_tx_disable(dev);
 
 	/* Disable MAC receive */

@@ -3455,7 +3457,6 @@ static void bcmgenet_netif_stop(struct net_device *dev, bool stop_phy)
 	/* Disable MAC transmit. TX DMA disabled must be done before this */
 	umac_enable_set(priv, CMD_TX_EN, false);
 
-	bcmgenet_disable_tx_napi(priv);
 	bcmgenet_disable_rx_napi(priv);
 	bcmgenet_intr_disable(priv);
 

-- 
2.53.0