Re: [PATCH 1/2] ipv4: initialize fib_trie prior to register_netdev_notifier call.

[ebiederm@xmission.com (Eric W. Biederman)]

Mahesh Bandewar <mahesh@bandewar.net> writes:

> From: Mahesh Bandewar <maheshb@google.com>
>
> Net stack initialization currently initializes fib-trie after the
> first call to netdevice_notifier() call. It does not cause any problem
> since there are no devices UP at this moment, but trying to bring 'lo'
> UP at initialization would make this assumption wrong. However changing
> order solves the issue.

This looks like a real issue and you are part of the way to a real fix.
The principle being you should not register things notifications until
you are ready to handle them.

As such fib_trie_init (which allocates the slabs) needs to come before
rtnl_register.  As a rtnl message can trigger slab allocation.

I have not traced it through but I suspect
register_pernet_subsys(&fib_net_ops) also needs to come before
rtnl_register.

Sigh.  It looks like this patch can be labeled:

Fixes: 7b1a74fdbb9e ("[NETNS]: Refactor fib initialization so it can handle multiple namespaces.")
Fixes: 7f9b80529b8a ("[IPV4]: fib hash|trie initialization")

So I really think the code needs to say:

void __init ip_fib_init(void)
{
	fib_trie_init();
	register_pernet_subsys(&fib_net_ops);

	register_netdevice_notifier(&fib_netdev_notifier);
	register_inetaddr_notifier(&fib_inetaddr_notifier);

	rtnl_register(PF_INET, RTM_NEWROUTE, inet_rtm_newroute, NULL, NULL);
	rtnl_register(PF_INET, RTM_DELROUTE, inet_rtm_delroute, NULL, NULL);
	rtnl_register(PF_INET, RTM_GETROUTE, NULL, inet_dump_fib, NULL);
}

Eric