]> git.ipfire.org Git - thirdparty/linux.git/commitdiff
net: do not acquire dev->tx_global_lock in netdev_watchdog_up()
authorEric Dumazet <edumazet@google.com>
Mon, 22 Jun 2026 11:01:08 +0000 (11:01 +0000)
committerJakub Kicinski <kuba@kernel.org>
Tue, 23 Jun 2026 21:40:33 +0000 (14:40 -0700)
Marek Szyprowski reported a deadlock during system resume when virtio_net
driver is used.

The deadlock occurs because netif_device_attach() is called while holding
dev->tx_global_lock (via netif_tx_lock_bh() in virtnet_restore_up()).
netif_device_attach() calls __netdev_watchdog_up(), which now also tries
to acquire dev->tx_global_lock to synchronize with dev_watchdog().

This recursive lock acquisition results in a deadlock.

Fix this by removing the tx_global_lock acquisition from netdev_watchdog_up().

The critical state (watchdog_timer and watchdog_ref_held) is already
protected by dev->watchdog_lock, which was introduced in the blamed commit.

Fixes: 8eed5519e496 ("net: watchdog: fix refcount tracking races")
Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
Closes: https://lore.kernel.org/netdev/a443376e-5187-4268-93b3-58047ef113a8@samsung.com/
Signed-off-by: Eric Dumazet <edumazet@google.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20260622110108.69541-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
net/sched/sch_generic.c

index 3f1c510df850dbdbaf10d483547c7b1f3a5d5482..ef2b4bf51564173751c74fefe17e3913ed2fa056 100644 (file)
@@ -594,9 +594,8 @@ void netdev_watchdog_up(struct net_device *dev)
                return;
        if (dev->watchdog_timeo <= 0)
                dev->watchdog_timeo = 5*HZ;
-       spin_lock_bh(&dev->tx_global_lock);
 
-       spin_lock(&dev->watchdog_lock);
+       spin_lock_bh(&dev->watchdog_lock);
        if (!mod_timer(&dev->watchdog_timer,
                       round_jiffies(jiffies + dev->watchdog_timeo))) {
                if (!dev->watchdog_ref_held) {
@@ -605,9 +604,7 @@ void netdev_watchdog_up(struct net_device *dev)
                        dev->watchdog_ref_held = true;
                }
        }
-       spin_unlock(&dev->watchdog_lock);
-
-       spin_unlock_bh(&dev->tx_global_lock);
+       spin_unlock_bh(&dev->watchdog_lock);
 }
 EXPORT_SYMBOL_GPL(netdev_watchdog_up);