net: refine dev_put()/dev_hold() debugging

We are still chasing some syzbot reports where we think a rogue dev_put() is called with no corresponding prior dev_hold(). Unfortunately it eats a reference on dev->dev_refcnt taken by innocent dev_hold_track(), meaning that the refcount saturation splat comes too late to be useful. Make sure that 'not tracked' dev_put() and dev_hold() better use CONFIG_NET_DEV_REFCNT_TRACKER=y debug infrastructure: Prior patch in the series allowed ref_tracker_alloc() and ref_tracker_free() to be called with a NULL @trackerp parameter, and to use a separate refcount only to detect too many put() even in the following case: dev_hold_track(dev, tracker_1, GFP_ATOMIC); dev_hold(dev); dev_put(dev); dev_put(dev); // Should complain loudly here. dev_put_track(dev, tracker_1); // instead of here Add clarification about netdev_tracker_alloc() role. v2: I replaced the dev_put() in linkwatch_do_dev() with __dev_put() because callers called netdev_tracker_free().

Eric Dumazet committed 4 years ago

Commit 4c6c11e1 parent 8fd55223 files · +50 −27Browse files

Unified · Split

3 files

include/linux/netdevice.h +46 −23
net/core/dev.c +1 −1
net/core/link_watch.c +3 −3

include/linux/netdevice.h+46−23

 /* Called by rtnetlink.c:rtnl_unlock() */
 void netdev_run_todo(void);
-/**
+static inline void __dev_put(struct net_device *dev)
- *	dev_put - release reference to device
- *	@dev: network device
- *
- * Release reference to device to allow it to be freed.
- * Try using dev_put_track() instead.
- */
-static inline void dev_put(struct net_device *dev)
 {
 	if (dev) {
 #ifdef CONFIG_PCPU_DEV_REFCNT
 	}
 }
-/**
+static inline void __dev_hold(struct net_device *dev)
- *	dev_hold - get reference to device
- *	@dev: network device
- *
- * Hold reference to device to keep it from being freed.
- * Try using dev_hold_track() instead.
- */
-static inline void dev_hold(struct net_device *dev)
 {
 	if (dev) {
 #ifdef CONFIG_PCPU_DEV_REFCNT
 	}
 }
+static inline void __netdev_tracker_alloc(struct net_device *dev,
+					  netdevice_tracker *tracker,
+					  gfp_t gfp)
+{
+#ifdef CONFIG_NET_DEV_REFCNT_TRACKER
+	ref_tracker_alloc(&dev->refcnt_tracker, tracker, gfp);
+#endif
+}
+/* netdev_tracker_alloc() can upgrade a prior untracked reference
+ * taken by dev_get_by_name()/dev_get_by_index() to a tracked one.
+ */
 static inline void netdev_tracker_alloc(struct net_device *dev,
 					netdevice_tracker *tracker, gfp_t gfp)
 {
 #ifdef CONFIG_NET_DEV_REFCNT_TRACKER
-	ref_tracker_alloc(&dev->refcnt_tracker, tracker, gfp);
+	refcount_dec(&dev->refcnt_tracker.no_tracker);
+	__netdev_tracker_alloc(dev, tracker, gfp);
 #endif
 }
 				  netdevice_tracker *tracker, gfp_t gfp)
 {
 	if (dev) {
-		dev_hold(dev);
+		__dev_hold(dev);
-		netdev_tracker_alloc(dev, tracker, gfp);
+		__netdev_tracker_alloc(dev, tracker, gfp);
 	}
 }
 {
 	if (dev) {
 		netdev_tracker_free(dev, tracker);
-		dev_put(dev);
+		__dev_put(dev);
 	}
 }
+/**
+ *	dev_hold - get reference to device
+ *	@dev: network device
+ *
+ * Hold reference to device to keep it from being freed.
+ * Try using dev_hold_track() instead.
+ */
+static inline void dev_hold(struct net_device *dev)
+{
+	dev_hold_track(dev, NULL, GFP_ATOMIC);
+}
+/**
+ *	dev_put - release reference to device
+ *	@dev: network device
+ *
+ * Release reference to device to allow it to be freed.
+ * Try using dev_put_track() instead.
+ */
+static inline void dev_put(struct net_device *dev)
+{
+	dev_put_track(dev, NULL);
+}
 static inline void dev_replace_track(struct net_device *odev,
 				     struct net_device *ndev,
 				     netdevice_tracker *tracker,
 	if (odev)
 		netdev_tracker_free(odev, tracker);
-	dev_hold(ndev);
+	__dev_hold(ndev);
-	dev_put(odev);
+	__dev_put(odev);
 	if (ndev)
-		netdev_tracker_alloc(ndev, tracker, gfp);
+		__netdev_tracker_alloc(ndev, tracker, gfp);
 }
 /* Carrier loss detection, dial on demand. The functions netif_carrier_on

net/core/dev.c+1−1

@@ -10172,7 +10172,7 @@ struct net_device alloc_netdev_mqs(int sizeof_priv, const char name, · 10171 unmodified lines
10172	dev->pcpu_refcnt = alloc_percpu(int);	10172	dev->pcpu_refcnt = alloc_percpu(int);
10173	if (!dev->pcpu_refcnt)	10173	if (!dev->pcpu_refcnt)
10174	goto free_dev;	10174	goto free_dev;
10175	dev_hold(dev);	10175	__dev_hold(dev);
10176	#else	10176	#else
10177	refcount_set(&dev->dev_refcnt, 1);	10177	refcount_set(&dev->dev_refcnt, 1);
10178	#endif	10178	#endif

net/core/link_watch.c+3−3

@@ -166,10 +166,10 @@ static void linkwatch_do_dev(struct net_device *dev) · 165 unmodified lines
166		166
167	netdev_state_change(dev);	167	netdev_state_change(dev);
168	}	168	}
169	/* Note: our callers are responsible for	169	/* Note: our callers are responsible for calling netdev_tracker_free().
170	* calling netdev_tracker_free().	170	* This is the reason we use __dev_put() instead of dev_put().
171	*/	171	*/
172	dev_put(dev);	172	__dev_put(dev);
173	}	173	}
174		174
175	static void __linkwatch_run_queue(int urgent_only)	175	static void __linkwatch_run_queue(int urgent_only)

net: refine dev_put()/dev_hold() debugging

Keyboard shortcuts

General

Issues and pull requests

Issues, pull requests and Inbox

Repository code