Am Freitag, 1. November 2013, 09:25:37 schrieb Patrick McHardy:

On Fri, Nov 01, 2013 at 09:45:29AM +0100, Steffen Klassert wrote:

quoted

On Wed, Oct 30, 2013 at 12:07:11AM +0000, Patrick McHardy wrote:

quoted

The problem is that the reassembled packet is referenced by the
individual
fragments, so we trigger the BUG_ON in pskb_expand_head(). In this
particular case the case we BUG() on is actually OK, but I'm looking at
a way we can fix this without special casing. Hope to have a patch for
testing in the next hours.

Just for the record. I'm observing similar, quite reproducable crashes
when
receiving fragmented icmp echo request packets on an IPsec gateway with
nf_conntrack_ipv6.

Since git commit 58a317f10 ("netfilter: ipv6: add IPv6 NAT support")
netfilter might insert a reassembled ipv6 packet with a shared skb and
local_df = 1 to the ok function. In case of xfrm, __xfrm6_output()
fragments the packet again and when adjusting the headroom later, we
crash because of a shared skb.

I can fix it by checking for a shared skb in ip6_fragment() and do
slow path fragmentation then. But we never needed such a check in
ip6_fragment(), so it's maybe better to fix it in netfilter.

So what seems to be happening is that this case in __ipv6_conntrack_in()
triggers:

        /* Conntrack helpers need the entire reassembled packet in the
         * POST_ROUTING hook. In case of unconfirmed connections NAT
         * might reassign a helper, so the entire packet is also
         * required.
         */
        ct = nf_ct_get(reasm, &ctinfo);
        if (ct != NULL && !nf_ct_is_untracked(ct)) {
                help = nfct_help(ct);
                if ((help && help->helper) || !nf_ct_is_confirmed(ct)) {
                        nf_conntrack_get_reasm(reasm);
                        NF_HOOK_THRESH(NFPROTO_IPV6, hooknum, reasm,
                                       (struct net_device *)in,
                                       (struct net_device *)out,
                                       okfn, NF_IP6_PRI_CONNTRACK + 1);

Since this code is called while walking through the fragment chain, we have
extra references to the reassembled skb. So I think what we need to do is
to release the fragment chain before calling NF_HOOK_THRESH() and indicate
this to nf_ct_frag6_output() so it will stop processing the chain
immediately.

I'll give it a try, will let you know when I have a patch for testing.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Are there patches available? I can crash my 3.12 kernel easily doing

	fping -p 20 -l -b 4000 bla

3.11.x does not expose this problem.

Regards,
-- 
Wolfgang Walter
Studentenwerk München
Anstalt des öffentlichen Rechts