Re: [PATCH] netfilter: fix dependency issues between IPv6 defragmentation and ip6tables

[liusdu <liusdu@126.com>]

On 2015年05月04日 04:09, Florian Westphal wrote:
> liusdu <liusdu@126.com> wrote:
>> On 2015年05月03日 19:07, Florian Westphal wrote:
>>> Liu Hua <sdu.liu@huawei.com> wrote:
>>>> commit f6318e558806c925029dc101f14874be9f9fa78f fix some related issue
>>>> when ip6tables is enabled. But when IP6_NF_IPTABLES is disabled and
>>>> NETFILTER_XT_TARGET_TPROXY is enabled. We will meet build failure with
>>>> "net/built-in.o: In function `tproxy_tg_init':
>>>> net/netfilter/xt_TPROXY.c:588: undefined reference to `nf_defrag_ipv6_enable'
>>>> "
>>>> So this patch change the Kconfig as ipv4 does.
>>>> --- a/net/netfilter/Kconfig
>>>> +++ b/net/netfilter/Kconfig
>>>> @@ -865,7 +865,7 @@ config NETFILTER_XT_TARGET_TPROXY
>>>>   	depends on (IPV6 || IPV6=n)
>>>>   	depends on IP_NF_MANGLE
>>>>   	select NF_DEFRAG_IPV4
>>>> -	select NF_DEFRAG_IPV6 if IP6_NF_IPTABLES
>>>> +	select NF_DEFRAG_IPV6
>>> IP6_NF_IPTABLES
>>> If IP6_NF_IPTABLES is not set, why would we have to pick
>>> up IPV6 defragmentation?
>>>
>>> Without ip6tables, TPROXY cannot be used for ipv6; in fact;
>>> xt_TPROXY should be built without ipv6 support in this case.
>>>
>>> My guess is that you have TPROXY=y but DEFRAG_IPV6=m, but that
>>> might warrant a better fix (xt_socket seems to have same issue).
>> Hi Florian,
>>
>> Yes, It was exactly what I did.  Actually there is a macro to
>> determine whether we compile nf_defrag_ipv6_enable or not,
>> called XT_TPROXY_HAVE_IPV6, which will be set to 1 while
>> IP6_NF_IPTABLES=y or m.  With the patch below  we can
>> make the compiling pass. But I am not sure it is good enough
>> or not.
>> diff --git a/net/netfilter/xt_TPROXY.c b/net/netfilter/xt_TPROXY.c
>> index cca96ce..abbda64 100644
>> --- a/net/netfilter/xt_TPROXY.c
>> +++ b/net/netfilter/xt_TPROXY.c
>> @@ -24,7 +24,7 @@
>>
>>   #include <net/netfilter/ipv4/nf_defrag_ipv4.h>
>>
>> -#if IS_ENABLED(CONFIG_IP6_NF_IPTABLES)
>> +#if IS_BUILTIN(CONFIG_IP6_NF_IPTABLES)
> This means you need to set IP6_NF_IPTABLES=y, else TPROXY doesn't work
> with ipv6.  I suggest to try and get Kconfig to force the needed
> dependency (DEFRAG=y if TPROXY=y && DEFRAG=m if TPROXY=m).
>
> Maybe this is enough (untested)
>
> diff --git a/net/netfilter/Kconfig b/net/netfilter/Kconfig
> --- a/net/netfilter/Kconfig
> +++ b/net/netfilter/Kconfig
> @@ -863,6 +863,7 @@ config NETFILTER_XT_TARGET_TPROXY
>          depends on NETFILTER_XTABLES
>          depends on NETFILTER_ADVANCED
>          depends on (IPV6 || IPV6=n)
> +       depends on (IP6_NF_IPTABLES || IP6_NF_IPTABLES=n)
>          depends on IP_NF_MANGLE
>          select NF_DEFRAG_IPV4
>          select NF_DEFRAG_IPV6 if IP6_NF_IPTABLES
> @@ -1356,6 +1357,7 @@ config NETFILTER_XT_MATCH_SOCKET
>          depends on NETFILTER_ADVANCED
>          depends on !NF_CONNTRACK || NF_CONNTRACK
>          depends on (IPV6 || IPV6=n)
> +       depends on (IP6_NF_IPTABLES || IP6_NF_IPTABLES=n)
>          select NF_DEFRAG_IPV4
>          select NF_DEFRAG_IPV6 if IP6_NF_IPTABLES
>          help
It works for me. With it DEFRAG, TPROXY and IP6_NF_IPTABLES are always  
compiled as
the same type. A better way to solve such dependency problem.

Since I only report this, and you solve it. So could you give a formal 
patch for this. I am a
beginner of linux network system and very glad to do the test for you.