Skip to contentSkip to conversation

Drop IPV6 packets if IPv6 is disabled on boot

PATCH In review Leonardo Bras 3 participants 17 messages Aug 30, 2019 – Sep 3, 2019

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

v4 cover Drop IPV6 packets if IPv6 is disabled on boot

This patchset was prevously a single patch named:
- netfilter: nf_tables: fib: Drop IPV6 packets if IPv6 is disabled on boot

It fixes a bug where a host, with IPv6 disabled on boot, has to deal with guest IPv6 packets, that comes from a bridge interface.
When these packets reach the host ip6tables they cause a kernel panic.

---
Changes from v3:
- Move drop logic from nft_fib6_eval{,_type} to nft_fib_netdev_eval
- Add another patch to drop ipv6 packets from bridge when ipv6 disabled

Changes from v2:
- Replace veredict.code from NF_DROP to NFT_BREAK
- Updated commit message (s/package/packet)

Changes from v1:
- Move drop logic from nft_fib_inet_eval() to nft_fib6_eval{,_type} so it can affect other usages of these functions.

Leonardo Bras (2):
netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is
disabled
net: br_netfiler_hooks: Drops IPv6 packets if IPv6 module is not
loaded

 net/bridge/br_netfilter_hooks.c | 2 ++
 net/netfilter/nft_fib_netdev.c  | 3 +++
 2 files changed, 5 insertions(+)
Signature · 1 line
2.20.1
#1 lore raw

PATCH v4 1/2 netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is disabled

If IPv6 is disabled on boot (ipv6.disable=1), but nft_fib_inet ends up dealing with a IPv6 packet, it causes a kernel panic in fib6_node_lookup_1(), crashing in bad_page_fault.

The panic is caused by trying to deference a very low address (0x38 in ppc64le), due to ipv6.fib6_main_tbl = NULL.

BUG: Kernel NULL pointer dereference at 0x00000038

The kernel panic was reproduced in a host that disabled IPv6 on boot and have to process guest packets (coming from a bridge) using it's ip6tables.

Terminate rule evaluation when packet protocol is IPv6 but the ipv6 module is not loaded.

---
 net/netfilter/nft_fib_netdev.c | 3 +++
 1 file changed, 3 insertions(+)
net/netfilter/nft_fib_netdev.c +3 -0
index 2cf3f32fe6d2..a2e726ae7f07 100644
@@ -14,6 +14,7 @@
 #include <linux/netfilter/nf_tables.h>
 #include <net/netfilter/nf_tables_core.h>
 #include <net/netfilter/nf_tables.h>
+#include <net/ipv6.h>
 
 #include <net/netfilter/nft_fib.h>
 
@@ -34,6 +35,8 @@ static void nft_fib_netdev_eval(const struct nft_expr *expr,
 		}
 		break;
 	case ETH_P_IPV6:
+		if (!ipv6_mod_enabled())
+			break;
 		switch (priv->result) {
 		case NFT_FIB_RESULT_OIF:
 		case NFT_FIB_RESULT_OIFNAME:
Signature · 1 line
2.20.1
maintainer Acked-by on v4 1/2
#2 lore raw

Re: [PATCH v4 1/2] netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is disabled

Quoted 8 lines · Leonardo Bras

If IPv6 is disabled on boot (ipv6.disable=1), but nft_fib_inet ends up dealing with a IPv6 packet, it causes a kernel panic in fib6_node_lookup_1(), crashing in bad_page_fault.

The panic is caused by trying to deference a very low address (0x38 in ppc64le), due to ipv6.fib6_main_tbl = NULL.

BUG: Kernel NULL pointer dereference at 0x00000038
Quoted 7 lines

The kernel panic was reproduced in a host that disabled IPv6 on boot and have to process guest packets (coming from a bridge) using it's ip6tables.

Terminate rule evaluation when packet protocol is IPv6 but the ipv6 module is not loaded.

#3 lore raw

Re: [PATCH v4 1/2] netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is disabled

Quoted 12 lines · Florian Westphal
Quoted 11 lines · Leonardo Bras

If IPv6 is disabled on boot (ipv6.disable=1), but nft_fib_inet ends up dealing with a IPv6 packet, it causes a kernel panic in fib6_node_lookup_1(), crashing in bad_page_fault.

The panic is caused by trying to deference a very low address (0x38 in ppc64le), due to ipv6.fib6_main_tbl = NULL.

BUG: Kernel NULL pointer dereference at 0x00000038

The kernel panic was reproduced in a host that disabled IPv6 on boot and have to process guest packets (coming from a bridge) using it's ip6tables.

Quoted 7 lines
Quoted 4 lines

Terminate rule evaluation when packet protocol is IPv6 but the ipv6 module is not loaded.

Hello Pablo,

Any trouble with this patch?
I could see the other* one got applied, but not this one. *(The other did not get acked, so i released it alone as v5)

Is there any fix I need to do in this one?

Best regards,
Leonardo Bras

#4 lore raw

Re: [PATCH v4 1/2] netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is disabled

Quoted 21 lines · Leonardo Bras
Quoted 12 lines · Florian Westphal
Quoted 11 lines · Leonardo Bras

If IPv6 is disabled on boot (ipv6.disable=1), but nft_fib_inet ends up dealing with a IPv6 packet, it causes a kernel panic in fib6_node_lookup_1(), crashing in bad_page_fault.

The panic is caused by trying to deference a very low address (0x38 in ppc64le), due to ipv6.fib6_main_tbl = NULL.

BUG: Kernel NULL pointer dereference at 0x00000038

The kernel panic was reproduced in a host that disabled IPv6 on boot and have to process guest packets (coming from a bridge) using it's ip6tables.

Quoted 7 lines
Quoted 4 lines

Terminate rule evaluation when packet protocol is IPv6 but the ipv6 module is not loaded.

Quoted 7 lines

Hello Pablo,

Any trouble with this patch?
I could see the other* one got applied, but not this one. *(The other did not get acked, so i released it alone as v5)

Is there any fix I need to do in this one?

Hm, I see, so this one:

https://patchwork.ozlabs.org/patch/1156100/

is not enough?

I was expecting we could find a way to handle this from br_netfilter alone itself.

Thanks.

#5 lore raw

Re: [PATCH v4 1/2] netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is disabled

Quoted 8 lines · Pablo Neira Ayuso
Quoted 7 lines · Leonardo Bras

On Fri, 2019-08-30 at 22:58 +0200, Florian Westphal wrote:
Hello Pablo,

Any trouble with this patch?
I could see the other* one got applied, but not this one. *(The other did not get acked, so i released it alone as v5)

Quoted 7 lines
Quoted 1 line

Is there any fix I need to do in this one?

Hm, I see, so this one:

https://patchwork.ozlabs.org/patch/1156100/

is not enough?

By what I could understand of Florian e-mail, we would need both:

Quoted 7 lines
Quoted 7 lines

So, given I don't want to plaster ipv6_mod_enabled() everywhere, I would suggest this course of action:

1. add a patch to BREAK in nft_fib_netdev.c for !ipv6_mod_enabled()
2. change net/bridge/br_netfilter_hooks.c, br_nf_pre_routing() to make sure ipv6_mod_enabled() is true before doing the ipv6 stack "emulation".

Is that ok?

Quoted 5 lines

I was expecting we could find a way to handle this from br_netfilter alone itself.

Thanks.

Thank you!

maintainer
#6 lore raw

Re: [PATCH v4 1/2] netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is disabled

In reply to Pablo Neira Ayuso
Quoted 28 lines · Pablo Neira Ayuso
Quoted 27 lines · Leonardo Bras
Quoted 12 lines · Florian Westphal
Quoted 11 lines · Leonardo Bras

If IPv6 is disabled on boot (ipv6.disable=1), but nft_fib_inet ends up dealing with a IPv6 packet, it causes a kernel panic in fib6_node_lookup_1(), crashing in bad_page_fault.

The panic is caused by trying to deference a very low address (0x38 in ppc64le), due to ipv6.fib6_main_tbl = NULL.

BUG: Kernel NULL pointer dereference at 0x00000038

The kernel panic was reproduced in a host that disabled IPv6 on boot and have to process guest packets (coming from a bridge) using it's ip6tables.

Quoted 7 lines
Quoted 4 lines

Terminate rule evaluation when packet protocol is IPv6 but the ipv6 module is not loaded.

Hello Pablo,

Any trouble with this patch?
I could see the other* one got applied, but not this one. *(The other did not get acked, so i released it alone as v5)

Quoted 7 lines
Quoted 1 line

Is there any fix I need to do in this one?

Hm, I see, so this one:

https://patchwork.ozlabs.org/patch/1156100/

is not enough?

No, its not.

Quoted 2 lines

I was expecting we could find a way to handle this from br_netfilter alone itself.

We can't because we support ipv6 fib lookups from the netdev family as well.

Alternative is to auto-accept ipv6 packets from the nf_tables eval loop, but I think its worse.

#7 lore raw

Re: [PATCH v4 1/2] netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is disabled

Quoted 39 lines · Florian Westphal
Quoted 28 lines · Pablo Neira Ayuso
Quoted 27 lines · Leonardo Bras
Quoted 12 lines · Florian Westphal
Quoted 11 lines · Leonardo Bras

If IPv6 is disabled on boot (ipv6.disable=1), but nft_fib_inet ends up dealing with a IPv6 packet, it causes a kernel panic in fib6_node_lookup_1(), crashing in bad_page_fault.

The panic is caused by trying to deference a very low address (0x38 in ppc64le), due to ipv6.fib6_main_tbl = NULL.

BUG: Kernel NULL pointer dereference at 0x00000038

The kernel panic was reproduced in a host that disabled IPv6 on boot and have to process guest packets (coming from a bridge) using it's ip6tables.

Quoted 7 lines
Quoted 4 lines

Terminate rule evaluation when packet protocol is IPv6 but the ipv6 module is not loaded.

Hello Pablo,

Any trouble with this patch?
I could see the other* one got applied, but not this one. *(The other did not get acked, so i released it alone as v5)

Quoted 7 lines
Quoted 1 line

Is there any fix I need to do in this one?

Hm, I see, so this one:

https://patchwork.ozlabs.org/patch/1156100/

is not enough?

No, its not.

Quoted 8 lines
Quoted 2 lines

I was expecting we could find a way to handle this from br_netfilter alone itself.

We can't because we support ipv6 fib lookups from the netdev family as well.

Alternative is to auto-accept ipv6 packets from the nf_tables eval loop, but I think its worse.

Could we add a restriction for nf_tables + br_netfilter + !ipv6. I mean, if this is an IPv6 packet, nf_tables is on and IPv6 module if off, then drop this packet?

By dropping packet, the user could diagnose that its setup is incomplete. I mean, if nf_tables fib ipv6 is used, then this setup is really wrong and the user forgots to load the ipv6 module.

maintainer
#8 lore raw

Re: [PATCH v4 1/2] netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is disabled

Quoted 6 lines · Pablo Neira Ayuso
Quoted 6 lines
Quoted 2 lines

I was expecting we could find a way to handle this from br_netfilter alone itself.

We can't because we support ipv6 fib lookups from the netdev family as well.

Quoted 6 lines
Quoted 2 lines

Alternative is to auto-accept ipv6 packets from the nf_tables eval loop, but I think its worse.

Could we add a restriction for nf_tables + br_netfilter + !ipv6. I mean, if this is an IPv6 packet, nf_tables is on and IPv6 module if off, then drop this packet?

We could do that from nft_do_chain_netdev().

#9 lore raw

Re: [PATCH v4 1/2] netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is disabled

Quoted 7 lines · Florian Westphal
Quoted 6 lines · Pablo Neira Ayuso
Quoted 6 lines
Quoted 2 lines

I was expecting we could find a way to handle this from br_netfilter alone itself.

We can't because we support ipv6 fib lookups from the netdev family as well.

Quoted 8 lines
Quoted 6 lines
Quoted 2 lines

Alternative is to auto-accept ipv6 packets from the nf_tables eval loop, but I think its worse.

Could we add a restriction for nf_tables + br_netfilter + !ipv6. I mean, if this is an IPv6 packet, nf_tables is on and IPv6 module if off, then drop this packet?

We could do that from nft_do_chain_netdev().

Indeed, this is all about the netdev case.

Probably add something similar to nf_ip6_route() to deal with ip6_route_lookup() case? This is the one trigering the problem, right?

BTW, how does nft_fib_ipv6 module kicks in if ipv6 module is not loaded? The symbol dependency would pull in the IPv6 module anyway.

maintainer
#10 lore raw

Re: [PATCH v4 1/2] netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is disabled

Quoted 7 lines · Pablo Neira Ayuso
Quoted 1 line · Florian Westphal

We could do that from nft_do_chain_netdev().

Indeed, this is all about the netdev case.

Probably add something similar to nf_ip6_route() to deal with ip6_route_lookup() case? This is the one trigering the problem, right?

Yes, this particular problem is caused by ipv6 fib not being initialized due to ipv6.disable=1. I don't know if there are cases other than FIB.

Quoted 2 lines

BTW, how does nft_fib_ipv6 module kicks in if ipv6 module is not loaded? The symbol dependency would pull in the IPv6 module anyway.

ipv6.disabled=1 does load the ipv6 module, but its non-functional.

#11 lore raw

Re: [PATCH v4 1/2] netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is disabled

Quoted 9 lines · Florian Westphal
Quoted 7 lines · Pablo Neira Ayuso
Quoted 1 line · Florian Westphal

We could do that from nft_do_chain_netdev().

Indeed, this is all about the netdev case.

Probably add something similar to nf_ip6_route() to deal with ip6_route_lookup() case? This is the one trigering the problem, right?

Quoted 8 lines

Yes, this particular problem is caused by ipv6 fib not being initialized due to ipv6.disable=1. I don't know if there are cases other than FIB.

Quoted 2 lines

BTW, how does nft_fib_ipv6 module kicks in if ipv6 module is not loaded? The symbol dependency would pull in the IPv6 module anyway.

ipv6.disabled=1 does load the ipv6 module, but its non-functional.

I see, thanks for explaining.

#12 lore raw

Re: [PATCH v4 1/2] netfilter: Terminate rule eval if protocol=IPv6 and ipv6 module is disabled

In reply to Leonardo Bras
Quoted 6 lines · Leonardo Bras

If IPv6 is disabled on boot (ipv6.disable=1), but nft_fib_inet ends up dealing with a IPv6 packet, it causes a kernel panic in fib6_node_lookup_1(), crashing in bad_page_fault.

The panic is caused by trying to deference a very low address (0x38 in ppc64le), due to ipv6.fib6_main_tbl = NULL.

Quoted 7 lines
BUG: Kernel NULL pointer dereference at 0x00000038

The kernel panic was reproduced in a host that disabled IPv6 on boot and have to process guest packets (coming from a bridge) using it's ip6tables.

Terminate rule evaluation when packet protocol is IPv6 but the ipv6 module is not loaded.

Patch is applied, thanks.

#13 lore raw

PATCH v4 2/2 net: br_netfiler_hooks: Drops IPv6 packets if IPv6 module is not loaded

In reply to Leonardo Bras

A kernel panic can happen if a host has disabled IPv6 on boot and have to process guest packets (coming from a bridge) using it's ip6tables.

IPv6 packets need to be dropped if the IPv6 module is not loaded.

---
 net/bridge/br_netfilter_hooks.c | 2 ++
 1 file changed, 2 insertions(+)
net/bridge/br_netfilter_hooks.c +2 -0
index d3f9592f4ff8..5e8693730df1 100644
@@ -493,6 +493,8 @@ static unsigned int br_nf_pre_routing(void *priv,
 	brnet = net_generic(state->net, brnf_net_id);
 	if (IS_IPV6(skb) || is_vlan_ipv6(skb, state->net) ||
 	    is_pppoe_ipv6(skb, state->net)) {
+		if (!ipv6_mod_enabled())
+			return NF_DROP;
 		if (!brnet->call_ip6tables &&
 		    !br_opt_get(br, BROPT_NF_CALL_IP6TABLES))
 			return NF_ACCEPT;
Signature · 1 line
2.20.1
maintainer
#14 lore raw

Re: [PATCH v4 2/2] net: br_netfiler_hooks: Drops IPv6 packets if IPv6 module is not loaded

Quoted 15 lines · Leonardo Bras

A kernel panic can happen if a host has disabled IPv6 on boot and have to process guest packets (coming from a bridge) using it's ip6tables.

IPv6 packets need to be dropped if the IPv6 module is not loaded.

---
 net/bridge/br_netfilter_hooks.c | 2 ++
 1 file changed, 2 insertions(+)
net/bridge/br_netfilter_hooks.c +0 -0
index d3f9592f4ff8..5e8693730df1 100644

net/bridge/br_netfilter_hooks.c static unsigned int br_nf_pre_routing(void *priv, · line 500

Quoted 8 lines
net/bridge/br_netfilter_hooks.c +2 -0
@@ -493,6 +493,8 @@ static unsigned int br_nf_pre_routing(void *priv,
 	brnet = net_generic(state->net, brnf_net_id);
 	if (IS_IPV6(skb) || is_vlan_ipv6(skb, state->net) ||
 	    is_pppoe_ipv6(skb, state->net)) {
+		if (!ipv6_mod_enabled())
+			return NF_DROP;
 		if (!brnet->call_ip6tables &&
 		    !br_opt_get(br, BROPT_NF_CALL_IP6TABLES))
 			return NF_ACCEPT;

No, thats too aggressive and turns the bridge into an ipv6 blackhole.

There are two solutions:
1. The above patch, but use NF_ACCEPT instead
2. keep the DROP, but move it below the call_ip6tables test, so that users can tweak call-ip6tables to accept packets.

Perhaps it would be good to also add a pr_warn_once() that tells that ipv6 was disabled on command line and call-ip6tables isn't supported in this configuration.

I would go with option two.

#15 lore raw

Re: [PATCH v4 2/2] net: br_netfiler_hooks: Drops IPv6 packets if IPv6 module is not loaded

Quoted 27 lines · Florian Westphal
Quoted 15 lines · Leonardo Bras

A kernel panic can happen if a host has disabled IPv6 on boot and have to process guest packets (coming from a bridge) using it's ip6tables.

IPv6 packets need to be dropped if the IPv6 module is not loaded.

---
 net/bridge/br_netfilter_hooks.c | 2 ++
 1 file changed, 2 insertions(+)
net/bridge/br_netfilter_hooks.c +0 -0
index d3f9592f4ff8..5e8693730df1 100644

net/bridge/br_netfilter_hooks.c static unsigned int br_nf_pre_routing(void *priv, · line 500

Quoted 8 lines
net/bridge/br_netfilter_hooks.c +2 -0
@@ -493,6 +493,8 @@ static unsigned int br_nf_pre_routing(void *priv,
 	brnet = net_generic(state->net, brnf_net_id);
 	if (IS_IPV6(skb) || is_vlan_ipv6(skb, state->net) ||
 	    is_pppoe_ipv6(skb, state->net)) {
+		if (!ipv6_mod_enabled())
+			return NF_DROP;
 		if (!brnet->call_ip6tables &&
 		    !br_opt_get(br, BROPT_NF_CALL_IP6TABLES))
 			return NF_ACCEPT;

No, thats too aggressive and turns the bridge into an ipv6 blackhole.

Quoted 4 lines

There are two solutions:
1. The above patch, but use NF_ACCEPT instead
2. keep the DROP, but move it below the call_ip6tables test, so that users can tweak call-ip6tables to accept packets.

Q: Does 2 mean that it will only be dropped if bridge intents to use host's ip6tables? Else, it will be accepted by previous if?

Quoted 4 lines

Perhaps it would be good to also add a pr_warn_once() that tells that ipv6 was disabled on command line and call-ip6tables isn't supported in this configuration.

Good idea, added.

Quoted 1 line

I would go with option two.

I think it's better than 1 too.

I sent a v5 with these changes:
https://lkml.org/lkml/2019/8/31/4

Thanks!

Leonardo Bras

maintainer
#16 lore raw

Re: [PATCH v4 2/2] net: br_netfiler_hooks: Drops IPv6 packets if IPv6 module is not loaded

Quoted 7 lines · Leonardo Bras
Quoted 4 lines

There are two solutions:
1. The above patch, but use NF_ACCEPT instead
2. keep the DROP, but move it below the call_ip6tables test, so that users can tweak call-ip6tables to accept packets.

Q: Does 2 mean that it will only be dropped if bridge intents to use host's ip6tables? Else, it will be accepted by previous if?

Yes, thats the idea: Let users decide if ipv6.disable or call-ip6tables is more important to them.