Re: [2.6.14-rt1] slowdown / oops.

[Ingo Molnar <mingo@elte.hu>]

more updates: NETFILTER_DEBUG catches the situation too - the problem 
seems to be wrong reference counts on the skb. I've also added a few 
printouts of ct fields:

NF_IP_ASSERT: include/linux/netfilter_ipv4/ip_conntrack.h:329(ip_conntrack_put)
NF_IP_ASSERT: net/ipv4/netfilter/ip_conntrack_core.c:346(destroy_conntrack)
NF_IP_ASSERT: net/ipv4/netfilter/ip_conntrack_core.c:346(destroy_conntrack)
BUG: NAT ct f0d6cf10 still hashed!
.. list empty ORIG: 0, REPLY: 0
.. expecting: 0
.. timeout pending: 1
.. ct_general: 0 [f0d6cf10]
.. nfct: f0d6cf10
 [<c0103d23>] dump_stack+0x13/0x20 (12)
 [<c03a878b>] check_hashes+0x13b/0x150 (48)
 [<c03a8878>] destroy_conntrack+0xd8/0x1f0 (32)
 [<c0373cf2>] ip_local_deliver+0x182/0x270 (40)
 [<c03742f5>] ip_rcv+0x2f5/0x4e0 (64)
 [<c035f95d>] netif_receive_skb+0x15d/0x1e0 (52)
 [<c02f2d97>] rtl8139_rx+0x1b7/0x340 (80)
 [<c02f3108>] rtl8139_poll+0x58/0x110 (40)
 [<c035fb32>] net_rx_action+0x72/0x140 (24)
 [<c011ee09>] ksoftirqd+0xb9/0x140 (40)
 [<c012d7d4>] kthread+0x94/0xa0 (28)
 [<c01010e9>] kernel_thread_helper+0x5/0xc (138035228)

i.e. we have ct->timeout still pending but the reference count is 0?  
That cannot be right.

	Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/