I mean that in kernel code: init_conntrack

["zhaoyao" <zhaoyao@xxxxxxxxxxxxxxxxxxx>]

netfilter，您好！
    multicast code and broadcast check has been removed. 
    So i want to know whether bc and mc will create conntrack?


Another question:
    kernel/net/ipv4/netfilter/ip_conntrack_core.c:
    init_conntrack:

    if (ip_conntrack_max &&
      atomic_read (&ip_conntrack_count) >= ip_conntrack_max)
    {
    }

    if it is smp, i think it can't protect ip_conntrack_count, some condition it will be larger than ip_conntrack_max.
	

　　　　　　　　致
礼！
 				

　　　　　　　　zhaoyao
　　　　　　　　zhaoyao@xxxxxxxxxxxxxxxxxxx
　　　　　　　　　　2004-01-04