Skip to contentSkip to conversation

net: init ingress queue

PATCH In review Changli Gao 2 participants 3 messages Dec 4, 2010

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 net: init ingress queue

The dev field of ingress queue is forgot to initialized, then NULL pointer dereference happens in qdisc_alloc().

[  303.019348] BUG: unable to handle kernel NULL pointer dereference at 0000000000000398
[  303.020068] IP: [<ffffffff81472aab>] qdisc_alloc+0x9b/0xc0
[  303.020068] PGD 3d637067 PUD 3d03f067 PMD 0 
[  303.020068] Oops: 0000 [#1] SMP 
[  303.020068] last sysfs file: /sys/kernel/uevent_seqnum
[  303.020068] CPU 0 
[  303.020068] Modules linked in: sch_ingress ipv6
[  303.020068] 
[  303.020068] Pid: 3138, comm: tc Not tainted 2.6.37-rc1+ #90 /VirtualBox
[  303.020068] RIP: 0010:[<ffffffff81472aab>]  [<ffffffff81472aab>] qdisc_alloc+0x9b/0xc0
[  303.020068] RSP: 0018:ffff88003daf1938  EFLAGS: 00010246
[  303.020068] RAX: ffff88003db01400 RBX: ffffffffa00612a0 RCX: 0000000000000000
[  303.020068] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffff88003db01600
[  303.020068] RBP: ffff88003daf1948 R08: 0000000000000000 R09: ffff88003db01400
[  303.020068] R10: ffff88003fbe9fe8 R11: dead000000200200 R12: ffff88003d37c600
[  303.020068] R13: 00000000fffffff1 R14: ffffffffa00612a0 R15: 00000000fffffff1
[  303.020068] FS:  00007f1b3575a700(0000) GS:ffff88003fc00000(0000) knlGS:0000000000000000
[  303.020068] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[  303.020068] CR2: 0000000000000398 CR3: 000000003d2a9000 CR4: 00000000000006f0
[  303.020068] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  303.020068] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[  303.020068] Process tc (pid: 3138, threadinfo ffff88003daf0000, task ffff88003d407500)
[  303.020068] Stack:
[  303.020068]  ffff88003e231000 ffff88003d6e1024 ffff88003daf19b8 ffffffff81474d68
[  303.020068]  0000000000000001 0000000000000000 ffff88003d37c600 ffff88003daf1a08
[  303.020068]  0073736572676e69 0000000000000000 ffff88003daf19b8 ffff88003d6e1010
[  303.020068] Call Trace:
[  303.020068]  [<ffffffff81474d68>] qdisc_create+0x78/0x350
[  303.020068]  [<ffffffff81476849>] tc_modify_qdisc+0x339/0x590
[  303.020068]  [<ffffffff8146622f>] rtnetlink_rcv_msg+0x16f/0x280
[  303.020068]  [<ffffffff814660c0>] ? rtnetlink_rcv_msg+0x0/0x280
[  303.020068]  [<ffffffff8147d6d9>] netlink_rcv_skb+0xa9/0xd0
[  303.020068]  [<ffffffff814660b0>] rtnetlink_rcv+0x20/0x30
[  303.020068]  [<ffffffff8147ce85>] netlink_unicast+0x2c5/0x2e0
[  303.020068]  [<ffffffff8147e195>] netlink_sendmsg+0x245/0x360
[  303.020068]  [<ffffffff81444874>] sock_sendmsg+0xe4/0x110
[  303.020068]  [<ffffffff810c5e19>] ? find_get_page+0x19/0x90
[  303.020068]  [<ffffffff810c716a>] ? filemap_fault+0xca/0x4b0
[  303.020068]  [<ffffffff810c6cd5>] ? unlock_page+0x25/0x30
[  303.020068]  [<ffffffff81444b6d>] ? move_addr_to_kernel+0x5d/0x60
[  303.020068]  [<ffffffff814500bd>] ? verify_iovec+0x7d/0xf0
[  303.020068]  [<ffffffff814465b5>] sys_sendmsg+0x1e5/0x330
[  303.020068]  [<ffffffff8147b5c8>] ? netlink_table_ungrab+0x28/0x30
[  303.020068]  [<ffffffff81516fac>] ? do_page_fault+0x1dc/0x4c0
[  303.020068]  [<ffffffff81444abb>] ? move_addr_to_user+0x9b/0xb0
[  303.020068]  [<ffffffff8112463b>] ? alloc_fd+0x4b/0x140
[  303.020068]  [<ffffffff81446394>] ? sys_recvmsg+0x44/0x80
[  303.020068]  [<ffffffff81002eab>] system_call_fastpath+0x16/0x1b
[  303.020068] Code: 8d 90 88 00 00 00 48 89 90 88 00 00 00 48 89 90 90 00 00 00 48 8b 53 28 48 89 10 48 8b 53 30 4c 89 60 68 48 89 50 08 49 8b 14 24 <48> 8b 92 98 03 00 00 65 ff 02 c7 40 40 01 00 00 00 5b 41 5c c9 
[  303.020068] RIP  [<ffffffff81472aab>] qdisc_alloc+0x9b/0xc0
[  303.020068]  RSP <ffff88003daf1938>
[  303.020068] CR2: 0000000000000398
[  303.135737] ---[ end trace 5b6e09a3328c82e4 ]---
---
 net/core/dev.c |    2 ++
 1 file changed, 2 insertions(+)
net/core/dev.c +2 -0
index cd24374..8083c68 100644
@@ -5577,6 +5577,8 @@ struct netdev_queue *dev_ingress_queue_create(struct net_device *dev)
 	queue = kzalloc(sizeof(*queue), GFP_KERNEL);
 	if (!queue)
 		return NULL;
+	netdev_queue_numa_node_write(queue, -1);
+	queue->dev = dev;
 	netdev_init_one_queue(dev, queue, NULL);
 	queue->qdisc = &noop_qdisc;
 	queue->qdisc_sleeping = &noop_qdisc;
#1 lore raw
Quoted 2 lines · Changli Gao

The dev field of ingress queue is forgot to initialized, then NULL pointer dereference happens in qdisc_alloc().

< deleted oops >

Quoted 9 lines
---
 net/core/dev.c |    2 ++
 1 file changed, 2 insertions(+)
net/core/dev.c +0 -0
index cd24374..8083c68 100644

net/core/dev.c struct netdev_queue *dev_ingress_queue_create(struct net_device *dev) · line 5584

Quoted 8 lines
net/core/dev.c +2 -0
@@ -5577,6 +5577,8 @@ struct netdev_queue *dev_ingress_queue_create(struct net_device *dev)
 	queue = kzalloc(sizeof(*queue), GFP_KERNEL);
 	if (!queue)
 		return NULL;
+	netdev_queue_numa_node_write(queue, -1);
+	queue->dev = dev;
 	netdev_init_one_queue(dev, queue, NULL);
 	queue->qdisc = &noop_qdisc;
 	queue->qdisc_sleeping = &noop_qdisc;

Hi Changli, thanks for bug report and patch.

IMHO there is no point to include this long Oops report in Changelog for a net-next-2.6 temporary problem, there wont be any bugzilla report.

Instead, you could say it is a followup patch for commits 1d24eb4815d1e0 and f2cd2d3e9b3ef960. Two or three lines with relevant information.

I suggest you submit following patch instead, as this seems cleaner to me ?

(factorize the two inits in netdev_init_one_queue())

net/core/dev.c +2 -5
index cd24374..36e10b4 100644
@@ -5113,7 +5113,6 @@ static int netif_alloc_netdev_queues(struct net_device *dev)
 {
 	unsigned int count = dev->num_tx_queues;
 	struct netdev_queue *tx;
-	int i;
 
 	BUG_ON(count < 1);
 
@@ -5125,10 +5124,6 @@ static int netif_alloc_netdev_queues(struct net_device *dev)
 	}
 	dev->_tx = tx;
 
-	for (i = 0; i < count; i++) {
-		netdev_queue_numa_node_write(&tx[i], -1);
-		tx[i].dev = dev;
-	}
 	return 0;
 }
 
@@ -5140,6 +5135,8 @@ static void netdev_init_one_queue(struct net_device *dev,
 	spin_lock_init(&queue->_xmit_lock);
 	netdev_set_xmit_lockdep_class(&queue->_xmit_lock, dev->type);
 	queue->xmit_lock_owner = -1;
+	netdev_queue_numa_node_write(queue, -1);
+	queue->dev = dev;
 }
 
 static void netdev_init_queues(struct net_device *dev)
#2 lore raw
Quoted 26 lines · Eric Dumazet
Quoted 2 lines · Changli Gao

The dev field of ingress queue is forgot to initialized, then NULL pointer dereference happens in qdisc_alloc().

< deleted oops >

Quoted 9 lines

---
 net/core/dev.c |    2 ++

 1 file changed, 2 insertions(+)
net/core/dev.c +0 -0
index cd24374..8083c68 100644

net/core/dev.c struct netdev_queue *dev_ingress_queue_create(struct net_device *dev) · line 5584

Quoted 8 lines
net/core/dev.c +2 -0
@@ -5577,6 +5577,8 @@ struct netdev_queue *dev_ingress_queue_create(struct net_device *dev)
       queue = kzalloc(sizeof(*queue), GFP_KERNEL);
       if (!queue)
               return NULL;
+     netdev_queue_numa_node_write(queue, -1);
+     queue->dev = dev;
       netdev_init_one_queue(dev, queue, NULL);
       queue->qdisc = &noop_qdisc;
       queue->qdisc_sleeping = &noop_qdisc;

Hi Changli, thanks for bug report and patch.

Quoted 3 lines

IMHO there is no point to include this long Oops report in Changelog for a net-next-2.6 temporary problem, there wont be any bugzilla report.

OK, Thanks.

Quoted 36 lines

Instead, you could say it is a followup patch for commits 1d24eb4815d1e0 and f2cd2d3e9b3ef960. Two or three lines with relevant information.

I suggest you submit following patch instead, as this seems cleaner to me ?

(factorize the two inits in netdev_init_one_queue())

net/core/dev.c +0 -5
index cd24374..36e10b4 100644
@@ -5113,7 +5113,6 @@ static int netif_alloc_netdev_queues(struct net_device *dev)
  {
        unsigned int count = dev->num_tx_queues;
        struct netdev_queue *tx;
-       int i;
 
        BUG_ON(count < 1);
 
@@ -5125,10 +5124,6 @@ static int netif_alloc_netdev_queues(struct net_device *dev)
        }
        dev->_tx = tx;
 
-       for (i = 0; i < count; i++) {
-               netdev_queue_numa_node_write(&tx[i], -1);
-               tx[i].dev = dev;
-       }
        return 0;
  }
 

net/core/dev.c static void netdev_init_one_queue(struct net_device *dev, · line 5142

Quoted 8 lines
net/core/dev.c +2 -0
@@ -5140,6 +5135,8 @@ static void netdev_init_one_queue(struct net_device *dev,
        spin_lock_init(&queue->_xmit_lock);
        netdev_set_xmit_lockdep_class(&queue->_xmit_lock, dev->type);
        queue->xmit_lock_owner = -1;
+       netdev_queue_numa_node_write(queue, -1);
+       queue->dev = dev;
  }
 
  static void netdev_init_queues(struct net_device *dev)

I thought about it before submitting. I am not sure if there are some users of txq->dev before netdev_init_one_queue().

Signature · 2 lines
Regards,
Changli Gao([email protected])