* [PATCH nf] ipvs: Fix panic due to non-linear skb
2014-05-26 1:37 [GIT PULL nf] Fixes for v3.15 Simon Horman
@ 2014-05-26 1:37 ` Simon Horman
2014-05-29 21:00 ` [GIT PULL nf] Fixes for v3.15 Pablo Neira Ayuso
1 sibling, 0 replies; 3+ messages in thread
From: Simon Horman @ 2014-05-26 1:37 UTC (permalink / raw)
To: Pablo Neira Ayuso
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov, Peter Christensen, Simon Horman
From: Peter Christensen <pch@ordbogen.com>
Receiving a ICMP response to an IPIP packet in a non-linear skb could
cause a kernel panic in __skb_pull.
The problem was introduced in
commit f2edb9f7706dcb2c0d9a362b2ba849efe3a97f5e ("ipvs: implement
passive PMTUD for IPIP packets").
Signed-off-by: Peter Christensen <pch@ordbogen.com>
Acked-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
---
net/netfilter/ipvs/ip_vs_core.c | 15 ++++++++++-----
1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/net/netfilter/ipvs/ip_vs_core.c b/net/netfilter/ipvs/ip_vs_core.c
index 4f26ee4..3d2d2c8 100644
--- a/net/netfilter/ipvs/ip_vs_core.c
+++ b/net/netfilter/ipvs/ip_vs_core.c
@@ -1392,15 +1392,19 @@ ip_vs_in_icmp(struct sk_buff *skb, int *related, unsigned int hooknum)
if (ipip) {
__be32 info = ic->un.gateway;
+ __u8 type = ic->type;
+ __u8 code = ic->code;
/* Update the MTU */
if (ic->type == ICMP_DEST_UNREACH &&
ic->code == ICMP_FRAG_NEEDED) {
struct ip_vs_dest *dest = cp->dest;
u32 mtu = ntohs(ic->un.frag.mtu);
+ __be16 frag_off = cih->frag_off;
/* Strip outer IP and ICMP, go to IPIP header */
- __skb_pull(skb, ihl + sizeof(_icmph));
+ if (pskb_pull(skb, ihl + sizeof(_icmph)) == NULL)
+ goto ignore_ipip;
offset2 -= ihl + sizeof(_icmph);
skb_reset_network_header(skb);
IP_VS_DBG(12, "ICMP for IPIP %pI4->%pI4: mtu=%u\n",
@@ -1408,7 +1412,7 @@ ip_vs_in_icmp(struct sk_buff *skb, int *related, unsigned int hooknum)
ipv4_update_pmtu(skb, dev_net(skb->dev),
mtu, 0, 0, 0, 0);
/* Client uses PMTUD? */
- if (!(cih->frag_off & htons(IP_DF)))
+ if (!(frag_off & htons(IP_DF)))
goto ignore_ipip;
/* Prefer the resulting PMTU */
if (dest) {
@@ -1427,12 +1431,13 @@ ip_vs_in_icmp(struct sk_buff *skb, int *related, unsigned int hooknum)
/* Strip outer IP, ICMP and IPIP, go to IP header of
* original request.
*/
- __skb_pull(skb, offset2);
+ if (pskb_pull(skb, offset2) == NULL)
+ goto ignore_ipip;
skb_reset_network_header(skb);
IP_VS_DBG(12, "Sending ICMP for %pI4->%pI4: t=%u, c=%u, i=%u\n",
&ip_hdr(skb)->saddr, &ip_hdr(skb)->daddr,
- ic->type, ic->code, ntohl(info));
- icmp_send(skb, ic->type, ic->code, info);
+ type, code, ntohl(info));
+ icmp_send(skb, type, code, info);
/* ICMP can be shorter but anyways, account it */
ip_vs_out_stats(cp, skb);
--
1.8.5.2
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [GIT PULL nf] Fixes for v3.15
2014-05-26 1:37 [GIT PULL nf] Fixes for v3.15 Simon Horman
2014-05-26 1:37 ` [PATCH nf] ipvs: Fix panic due to non-linear skb Simon Horman
@ 2014-05-29 21:00 ` Pablo Neira Ayuso
1 sibling, 0 replies; 3+ messages in thread
From: Pablo Neira Ayuso @ 2014-05-29 21:00 UTC (permalink / raw)
To: Simon Horman
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov
On Mon, May 26, 2014 at 10:37:14AM +0900, Simon Horman wrote:
> Hi Pablo,
>
> I realise this is extremely late in the v3.15 cycle so please don't hesitate
> to push it back to v3.16. I am more than happy to rebase/resubmit if that
> is what you decided on.
>
> This fix resolves a panic due to non-linear skb.
>
> It has been present since v3.6-rc1.
>
> I would like this fix considered for -stable and I have checked that it
> applies cleanly on top of v3.10.40, v3.12.20 and v3.14.4.
>
>
> The following changes since commit 3b084e99a3fabaeb0f9c65a0806cde30f0b2835e:
>
> netfilter: nf_tables: fix trace of matching non-terminal rule (2014-05-15 19:44:20 +0200)
>
> are available in the git repository at:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/horms/ipvs.git tags/ipvs-fixes-for-v3.15
Pulled, thanks Simon.
^ permalink raw reply [flat|nested] 3+ messages in thread