Re: [PATCH v2 1/1] netfilter: nf_tables: fib: Drop IPV6 packages if IPv6 is disabled on boot

From: Leonardo Bras <leonardo@linux.ibm.com>
To: Pablo Neira Ayuso <pablo@netfilter.org>
Cc: netfilter-devel@vger.kernel.org, coreteam@netfilter.org,
	netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
	Jozsef Kadlecsik <kadlec@netfilter.org>,
	Florian Westphal <fw@strlen.de>,
	"David S. Miller" <davem@davemloft.net>,
	Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>,
	Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Subject: Re: [PATCH v2 1/1] netfilter: nf_tables: fib: Drop IPV6 packages if IPv6 is disabled on boot
Date: Tue, 27 Aug 2019 14:34:14 -0300	[thread overview]
Message-ID: <77c43754ff72e9a2e8048ccd032351cf0186080a.camel@linux.ibm.com> (raw)
In-Reply-To: <20190827103541.vzwqwg4jlbuzajxu@salvia>

[-- Attachment #1: Type: text/plain, Size: 1912 bytes --]

On Tue, 2019-08-27 at 12:35 +0200, Pablo Neira Ayuso wrote:
> On Wed, Aug 21, 2019 at 11:15:06AM -0300, Leonardo Bras wrote:
> > If IPv6 is disabled on boot (ipv6.disable=1), but nft_fib_inet ends up
> > dealing with a IPv6 package, it causes a kernel panic in
> > fib6_node_lookup_1(), crashing in bad_page_fault.
> 
> Q: How do you get to see IPv6 packets if IPv6 module is disable?
I could reproduce this bug on a host ('ipv6.disable=1') starting a
guest with a virtio-net interface with 'filterref' over a virtual
bridge. It crashes the host during guest boot (just before login).

By that I could understand that a guest IPv6 network traffic (viavirtio-net) may cause this kernel panic. 

> 
> > 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
> > 
> > Fix this behavior by dropping IPv6 packages if !ipv6_mod_enabled().
> 
> I'd suggest: s/package/packet/
Sure, I will make sure to put it on v3.
(Sorry, I am not very used to net subsystem.)
> 
> [...]
> > diff --git a/net/ipv6/netfilter/nft_fib_ipv6.c b/net/ipv6/netfilter/nft_fib_ipv6.c
> > index 7ece86afd079..75acc417e2ff 100644
> > --- a/net/ipv6/netfilter/nft_fib_ipv6.c
> > +++ b/net/ipv6/netfilter/nft_fib_ipv6.c
> > @@ -125,6 +125,11 @@ void nft_fib6_eval_type(const struct nft_expr *expr, struct nft_regs *regs,
> >  	u32 *dest = &regs->data[priv->dreg];
> >  	struct ipv6hdr *iph, _iph;
> >  
> > +	if (!ipv6_mod_enabled()) {
> > +		regs->verdict.code = NF_DROP;
> 
> NFT_BREAK instead to stop evaluating this rule, this results in a
> mismatch, so you let the user decide what to do with packets that do
> not match your policy.
Ok, I will replace for v3.

> 
> The drop case at the bottom of the fib eval function never actually
> never happens.
Which one do you mean?

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 833 bytes --]