hsr: fix a race condition in node list insertion and deletion

hsr nodes are protected by RCU and there is no write side lock.
But node insertions and deletions could be being operated concurrently.
So write side locking is needed.

Test commands:
    ip netns add nst
    ip link add veth0 type veth peer name veth1
    ip link add veth2 type veth peer name veth3
    ip link set veth1 netns nst
    ip link set veth3 netns nst
    ip link set veth0 up
    ip link set veth2 up
    ip link add hsr0 type hsr slave1 veth0 slave2 veth2
    ip a a 192.168.100.1/24 dev hsr0
    ip link set hsr0 up
    ip netns exec nst ip link set veth1 up
    ip netns exec nst ip link set veth3 up
    ip netns exec nst ip link add hsr1 type hsr slave1 veth1 slave2 veth3
    ip netns exec nst ip a a 192.168.100.2/24 dev hsr1
    ip netns exec nst ip link set hsr1 up

    for i in {0..9}
    do
        for j in {0..9}
	do
	    for k in {0..9}
	    do
	        for l in {0..9}
		do
	        arping 192.168.100.2 -I hsr0 -s 00:01:3$i:4$j:5$k:6$l -c1 &
		done
	    done
	done
    done

Splat looks like:
[  236.066091][ T3286] list_add corruption. next->prev should be prev (ffff8880a5940300), but was ffff8880a5940d0.
[  236.069617][ T3286] ------------[ cut here ]------------
[  236.070545][ T3286] kernel BUG at lib/list_debug.c:25!
[  236.071391][ T3286] invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN PTI
[  236.072343][ T3286] CPU: 0 PID: 3286 Comm: arping Tainted: G        W         5.5.0-rc1+ #209
[  236.073463][ T3286] Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
[  236.074695][ T3286] RIP: 0010:__list_add_valid+0x74/0xd0
[  236.075499][ T3286] Code: 48 39 da 75 27 48 39 f5 74 36 48 39 dd 74 31 48 83 c4 08 b8 01 00 00 00 5b 5d c3 48 b
[  236.078277][ T3286] RSP: 0018:ffff8880aaa97648 EFLAGS: 00010286
[  236.086991][ T3286] RAX: 0000000000000075 RBX: ffff8880d4624c20 RCX: 0000000000000000
[  236.088000][ T3286] RDX: 0000000000000075 RSI: 0000000000000008 RDI: ffffed1015552ebf
[  236.098897][ T3286] RBP: ffff88809b53d200 R08: ffffed101b3c04f9 R09: ffffed101b3c04f9
[  236.099960][ T3286] R10: 00000000308769a1 R11: ffffed101b3c04f8 R12: ffff8880d4624c28
[  236.100974][ T3286] R13: ffff8880d4624c20 R14: 0000000040310100 R15: ffff8880ce17ee02
[  236.138967][ T3286] FS:  00007f23479fa680(0000) GS:ffff8880d9c00000(0000) knlGS:0000000000000000
[  236.144852][ T3286] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  236.145720][ T3286] CR2: 00007f4a14bab210 CR3: 00000000a61c6001 CR4: 00000000000606f0
[  236.146776][ T3286] Call Trace:
[  236.147222][ T3286]  hsr_add_node+0x314/0x490 [hsr]
[  236.153633][ T3286]  hsr_forward_skb+0x2b6/0x1bc0 [hsr]
[  236.154362][ T3286]  ? rcu_read_lock_sched_held+0x90/0xc0
[  236.155091][ T3286]  ? rcu_read_lock_bh_held+0xa0/0xa0
[  236.156607][ T3286]  hsr_dev_xmit+0x70/0xd0 [hsr]
[  236.157254][ T3286]  dev_hard_start_xmit+0x160/0x740
[  236.157941][ T3286]  __dev_queue_xmit+0x1961/0x2e10
[  236.158565][ T3286]  ? netdev_core_pick_tx+0x2e0/0x2e0
[ ... ]

Reported-by: syzbot+3924327f9ad5f4d2b343@syzkaller.appspotmail.com
Fixes: f421436a591d ("net/hsr: Add support for the High-availability Seamless Redundancy protocol (HSRv0)")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 46 insertions, 27 deletions

diff --git a/net/hsr/hsr_framereg.c b/net/hsr/hsr_framereg.c
index 292be446007b..27dc65d7de67 100644
--- a/net/hsr/hsr_framereg.c
+++ b/net/hsr/hsr_framereg.c
@@ -75,10 +75,11 @@ static struct hsr_node *find_node_by_addr_A(struct list_head *node_db,
 /* Helper for device init; the self_node_db is used in hsr_rcv() to recognize
  * frames from self that's been looped over the HSR ring.
  */
-int hsr_create_self_node(struct list_head *self_node_db,
+int hsr_create_self_node(struct hsr_priv *hsr,
 			 unsigned char addr_a[ETH_ALEN],
 			 unsigned char addr_b[ETH_ALEN])
 {
+	struct list_head *self_node_db = &hsr->self_node_db;
 	struct hsr_node *node, *oldnode;
 
 	node = kmalloc(sizeof(*node), GFP_KERNEL);
@@ -88,33 +89,33 @@ int hsr_create_self_node(struct list_head *self_node_db,
 	ether_addr_copy(node->macaddress_A, addr_a);
 	ether_addr_copy(node->macaddress_B, addr_b);
 
-	rcu_read_lock();
+	spin_lock_bh(&hsr->list_lock);
 	oldnode = list_first_or_null_rcu(self_node_db,
 					 struct hsr_node, mac_list);
 	if (oldnode) {
 		list_replace_rcu(&oldnode->mac_list, &node->mac_list);
-		rcu_read_unlock();
-		synchronize_rcu();
-		kfree(oldnode);
+		spin_unlock_bh(&hsr->list_lock);
+		kfree_rcu(oldnode, rcu_head);
 	} else {
-		rcu_read_unlock();
 		list_add_tail_rcu(&node->mac_list, self_node_db);
+		spin_unlock_bh(&hsr->list_lock);
 	}
 
 	return 0;
 }
 
-void hsr_del_self_node(struct list_head *self_node_db)
+void hsr_del_self_node(struct hsr_priv *hsr)
 {
+	struct list_head *self_node_db = &hsr->self_node_db;
 	struct hsr_node *node;
 
-	rcu_read_lock();
+	spin_lock_bh(&hsr->list_lock);
 	node = list_first_or_null_rcu(self_node_db, struct hsr_node, mac_list);
-	rcu_read_unlock();
 	if (node) {
 		list_del_rcu(&node->mac_list);
-		kfree(node);
+		kfree_rcu(node, rcu_head);
 	}
+	spin_unlock_bh(&hsr->list_lock);
 }
 
 void hsr_del_nodes(struct list_head *node_db)
@@ -130,30 +131,43 @@ void hsr_del_nodes(struct list_head *node_db)
  * seq_out is used to initialize filtering of outgoing duplicate frames
  * originating from the newly added node.
  */
-struct hsr_node *hsr_add_node(struct list_head *node_db, unsigned char addr[],
-			      u16 seq_out)
+static struct hsr_node *hsr_add_node(struct hsr_priv *hsr,
+				     struct list_head *node_db,
+				     unsigned char addr[],
+				     u16 seq_out)
 {
-	struct hsr_node *node;
+	struct hsr_node *new_node, *node;
 	unsigned long now;
 	int i;
 
-	node = kzalloc(sizeof(*node), GFP_ATOMIC);
-	if (!node)
+	new_node = kzalloc(sizeof(*new_node), GFP_ATOMIC);
+	if (!new_node)
 		return NULL;
 
-	ether_addr_copy(node->macaddress_A, addr);
+	ether_addr_copy(new_node->macaddress_A, addr);
 
 	/* We are only interested in time diffs here, so use current jiffies
 	 * as initialization. (0 could trigger an spurious ring error warning).
 	 */
 	now = jiffies;
 	for (i = 0; i < HSR_PT_PORTS; i++)
-		node->time_in[i] = now;
+		new_node->time_in[i] = now;
 	for (i = 0; i < HSR_PT_PORTS; i++)
-		node->seq_out[i] = seq_out;
-
-	list_add_tail_rcu(&node->mac_list, node_db);
+		new_node->seq_out[i] = seq_out;
 
+	spin_lock_bh(&hsr->list_lock);
+	list_for_each_entry_rcu(node, node_db, mac_list) {
+		if (ether_addr_equal(node->macaddress_A, addr))
+			goto out;
+		if (ether_addr_equal(node->macaddress_B, addr))
+			goto out;
+	}
+	list_add_tail_rcu(&new_node->mac_list, node_db);
+	spin_unlock_bh(&hsr->list_lock);
+	return new_node;
+out:
+	spin_unlock_bh(&hsr->list_lock);
+	kfree(new_node);
 	return node;
 }
 
@@ -163,6 +177,7 @@ struct hsr_node *hsr_get_node(struct hsr_port *port, struct sk_buff *skb,
 			      bool is_sup)
 {
 	struct list_head *node_db = &port->hsr->node_db;
+	struct hsr_priv *hsr = port->hsr;
 	struct hsr_node *node;
 	struct ethhdr *ethhdr;
 	u16 seq_out;
@@ -196,7 +211,7 @@ struct hsr_node *hsr_get_node(struct hsr_port *port, struct sk_buff *skb,
 		seq_out = HSR_SEQNR_START;
 	}
 
-	return hsr_add_node(node_db, ethhdr->h_source, seq_out);
+	return hsr_add_node(hsr, node_db, ethhdr->h_source, seq_out);
 }
 
 /* Use the Supervision frame's info about an eventual macaddress_B for merging
@@ -206,10 +221,11 @@ struct hsr_node *hsr_get_node(struct hsr_port *port, struct sk_buff *skb,
 void hsr_handle_sup_frame(struct sk_buff *skb, struct hsr_node *node_curr,
 			  struct hsr_port *port_rcv)
 {
-	struct ethhdr *ethhdr;
-	struct hsr_node *node_real;
+	struct hsr_priv *hsr = port_rcv->hsr;
 	struct hsr_sup_payload *hsr_sp;
+	struct hsr_node *node_real;
 	struct list_head *node_db;
+	struct ethhdr *ethhdr;
 	int i;
 
 	ethhdr = (struct ethhdr *)skb_mac_header(skb);
@@ -231,7 +247,7 @@ void hsr_handle_sup_frame(struct sk_buff *skb, struct hsr_node *node_curr,
 	node_real = find_node_by_addr_A(node_db, hsr_sp->macaddress_A);
 	if (!node_real)
 		/* No frame received from AddrA of this node yet */
-		node_real = hsr_add_node(node_db, hsr_sp->macaddress_A,
+		node_real = hsr_add_node(hsr, node_db, hsr_sp->macaddress_A,
 					 HSR_SEQNR_START - 1);
 	if (!node_real)
 		goto done; /* No mem */
@@ -252,7 +268,9 @@ void hsr_handle_sup_frame(struct sk_buff *skb, struct hsr_node *node_curr,
 	}
 	node_real->addr_B_port = port_rcv->type;
 
+	spin_lock_bh(&hsr->list_lock);
 	list_del_rcu(&node_curr->mac_list);
+	spin_unlock_bh(&hsr->list_lock);
 	kfree_rcu(node_curr, rcu_head);
 
 done:
@@ -368,12 +386,13 @@ void hsr_prune_nodes(struct timer_list *t)
 {
 	struct hsr_priv *hsr = from_timer(hsr, t, prune_timer);
 	struct hsr_node *node;
+	struct hsr_node *tmp;
 	struct hsr_port *port;
 	unsigned long timestamp;
 	unsigned long time_a, time_b;
 
-	rcu_read_lock();
-	list_for_each_entry_rcu(node, &hsr->node_db, mac_list) {
+	spin_lock_bh(&hsr->list_lock);
+	list_for_each_entry_safe(node, tmp, &hsr->node_db, mac_list) {
 		/* Don't prune own node. Neither time_in[HSR_PT_SLAVE_A]
 		 * nor time_in[HSR_PT_SLAVE_B], will ever be updated for
 		 * the master port. Thus the master node will be repeatedly
@@ -421,7 +440,7 @@ void hsr_prune_nodes(struct timer_list *t)
 			kfree_rcu(node, rcu_head);
 		}
 	}
-	rcu_read_unlock();
+	spin_unlock_bh(&hsr->list_lock);
 
 	/* Restart timer */
 	mod_timer(&hsr->prune_timer,