Re: [PATCH] netfilter: Fix build failure when ipv6 but xt_tproxy is built in

From: Steven Rostedt <rostedt@goodmis.org>
Date: Tue, 29 Mar 2011 22:13:19 -0400

While running ktest.pl doing ranconfigs, the following build error
occurred:

Is sending netfilter patches to the explicitly listed maintainer in
MAINTAINERS too much to ask Stephen? :-/

CC:'d

On Tue, 2011-03-29 at 22:35 -0700, David Miller wrote:

From: Steven Rostedt <rostedt@goodmis.org>
Date: Tue, 29 Mar 2011 22:13:19 -0400

quoted

While running ktest.pl doing ranconfigs, the following build error
occurred:

Is sending netfilter patches to the explicitly listed maintainer in
MAINTAINERS too much to ask Stephen? :-/

CC:'d

Sorry David, I just did a git blame and Cc'd the ones that did the
original change. I'll do better next time.

-- Steve

On 30.03.2011 07:35, David Miller wrote:

From: Steven Rostedt <rostedt@goodmis.org>
Date: Tue, 29 Mar 2011 22:13:19 -0400

quoted

While running ktest.pl doing ranconfigs, the following build error
occurred:

Is sending netfilter patches to the explicitly listed maintainer in
MAINTAINERS too much to ask Stephen? :-/

CC:'d

quoted

net/built-in.o: In function `tproxy_tg6_v1':
/home/rostedt/work/autotest/nobackup/linux-test.git/net/netfilter/xt_TPROXY.c:288: undefined reference to `ipv6_find_hdr'

This happened because the xt_TPROXY code was compiled into the kernel
proper, but the ipv6 netfilter was compiled as a module. The fix is to
only enter the code that calls ipv6_find_hdr if ipv6 netfilter is
compiled into the kernel, or if it is a module, so is the xt_TPROXY
code.

I don't think this is a good fix for the problem since it may lead to
the confusing situation that both TPROXY and ip6tables are enabled,
but TPROXY has no IPv6 support.

I think we should solve this by either adding a Kconfig dependency
on (IP6_NF_IPTABLES || IP6_NF_IPTABLES=n) or using ipv6_skip_exthdr()
instead of ipv6_find_hdr().

Krisztian, what do you think?

quoted

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>

diff --git a/net/netfilter/xt_TPROXY.c b/net/netfilter/xt_TPROXY.c
index dcfd57e..5915f94 100644
--- a/net/netfilter/xt_TPROXY.c
+++ b/net/netfilter/xt_TPROXY.c

@@ -22,7 +22,13 @@
 
 #include <net/netfilter/ipv4/nf_defrag_ipv4.h>
 
-#if defined(CONFIG_IP6_NF_IPTABLES) || defined(CONFIG_IP6_NF_IPTABLES_MODULE)
+/*
+ * Only include the following if ip6 tables is compiled in
+ * the kernel, or it is a module and this code is also a module.
+ */
+#if defined(CONFIG_IP6_NF_IPTABLES) || \
+	(defined(CONFIG_IP6_NF_IPTABLES_MODULE) && \
+	 defined(CONFIG_NETFILTER_XT_TARGET_TPROXY_MODULE))
 #define XT_TPROXY_HAVE_IPV6 1
 #include <net/if_inet6.h>
 #include <net/addrconf.h>

Hi,

On Mon, 2011-04-04 at 15:54 +0200, Patrick McHardy wrote:

quoted

quoted

net/built-in.o: In function `tproxy_tg6_v1':
/home/rostedt/work/autotest/nobackup/linux-test.git/net/netfilter/xt_TPROXY.c:288: undefined reference to `ipv6_find_hdr'

This happened because the xt_TPROXY code was compiled into the kernel
proper, but the ipv6 netfilter was compiled as a module. The fix is to
only enter the code that calls ipv6_find_hdr if ipv6 netfilter is
compiled into the kernel, or if it is a module, so is the xt_TPROXY
code.

I don't think this is a good fix for the problem since it may lead to
the confusing situation that both TPROXY and ip6tables are enabled,
but TPROXY has no IPv6 support.

I think we should solve this by either adding a Kconfig dependency
on (IP6_NF_IPTABLES || IP6_NF_IPTABLES=n) or using ipv6_skip_exthdr()
instead of ipv6_find_hdr().

Krisztian, what do you think?

I'd definitely prefer using ipv6_skip_exthdr() instead of playing
various tricks with the config preprocessor macros.

What about something like this?

diff --git a/net/netfilter/xt_TPROXY.c b/net/netfilter/xt_TPROXY.c
index dcfd57e..1ef0e56 100644
--- a/net/netfilter/xt_TPROXY.c
+++ b/net/netfilter/xt_TPROXY.c

@@ -283,10 +283,10 @@ tproxy_tg6_v1(struct sk_buff *skb, const struct xt_action_param *par)
        const struct in6_addr *laddr;
        __be16 lport;
        int thoff;
-       int tproto;
+       u8 tproto = iph->nexthdr;
 
-       tproto = ipv6_find_hdr(skb, &thoff, -1, NULL);
-       if (tproto < 0) {
+       thoff = ipv6_skip_exthdr(skb, sizeof(*iph), &tproto);
+       if (thoff < 0) {
                pr_debug("unable to find transport header in IPv6 packet, dropping\n");
                return NF_DROP;
        }

--

KOVACS Krisztian

Am 05.04.2011 16:43, schrieb KOVACS Krisztian:

Looks good to me. Please formally submit this including a Signed-off-by:
line and I'll push it upstream.

Hi,

On 04/05/2011 04:49 PM, Patrick McHardy wrote:

quoted

What about something like this?

diff --git a/net/netfilter/xt_TPROXY.c b/net/netfilter/xt_TPROXY.c
index dcfd57e..1ef0e56 100644
--- a/net/netfilter/xt_TPROXY.c
+++ b/net/netfilter/xt_TPROXY.c

@@ -283,10 +283,10 @@ tproxy_tg6_v1(struct sk_buff *skb, const struct xt_action_param *par)
        const struct in6_addr *laddr;
        __be16 lport;
        int thoff;
-       int tproto;
+       u8 tproto = iph->nexthdr;
 
-       tproto = ipv6_find_hdr(skb, &thoff, -1, NULL);
-       if (tproto < 0) {
+       thoff = ipv6_skip_exthdr(skb, sizeof(*iph), &tproto);
+       if (thoff < 0) {
                pr_debug("unable to find transport header in IPv6 packet, dropping\n");
                return NF_DROP;
        }

Looks good to me. Please formally submit this including a Signed-off-by:
line and I'll push it upstream.

Will do, as soon as I have a chance to run the test suite. Thanks Patrick.

-- 
KOVACS Krisztian