Yes， I check the ipv6_stub->ipv6_fragment. And in the case
if there is no ipv6_stub->ipv6_fragment it means no ipv6 fragment support
and it should free the skb.
Maybe sometimes not all the calling remember to check with this? So
it should add a default one with following?
+static int eafnosupport_ipv6_fragment(struct net *net, struct sock *sk, struct
sk_buff *skb,
+ int (*output)(struct net *, struct sock
*, struct sk_buff *))
+{
+ kfree_skb(skb);
+ return -EAFNOSUPPORT;
+}
+
在 2020/8/28 6:38, Marcelo Ricardo Leitner 写道:
> On Thu, Aug 27, 2020 at 07:51:47AM -0700, David Miller wrote:
>> From: we...@ucloud.cn
>> Date: Thu, 27 Aug 2020 18:39:51 +0800
>>
>>> From: wenxu <we...@ucloud.cn>
>>>
>>> Add ipv6_fragment to ipv6_stub to avoid calling netfilter when
>>> access ip6_fragment.
>>>
>>> Signed-off-by: wenxu <we...@ucloud.cn>
>> Please test these changes with ipv6 disabled.
>>
>> It will crash, you have to update the default stub in
>> net/ipv6/addrconf_core.c as well.
> I didn't test it myself but I'm not seeing how the crash could happen.
> The next patch does check for it being NULL before using it:
>
> - if (!v6ops)
> + if (!ipv6_stub->ipv6_fragment)
> goto err;
>
> Wenxu?
>
> Marcelo
>
