[PATCH] xfrm: fix xfrm_input/xfrm_tunnel_check oops

[PATCH] xfrm: fix xfrm_input/xfrm_tunnel_check oops

[Alexey Dobriyan]

https://bugzilla.kernel.org/show_bug.cgi?id=95211

Commit 70be6c91c86596ad2b60c73587880b47df170a41
("xfrm: Add xfrm_tunnel_skb_cb to the skb common buffer") added check
which dereferences ->outer_mode too early but larval SAs don't have
this pointer set (yet). So check for tunnel stuff later.

Mike Noordermeer reported this bug and patiently applied all the debugging.

Technically this is remote-oops-in-interrupt-context type of thing.

BUG: unable to handle kernel NULL pointer dereference at 0000000000000034
IP: [<ffffffff8150dca2>] xfrm_input+0x3c2/0x5a0
	...
[<ffffffff81500fc6>] ? xfrm4_esp_rcv+0x36/0x70
[<ffffffff814acc9a>] ? ip_local_deliver_finish+0x9a/0x200
[<ffffffff81471b83>] ? __netif_receive_skb_core+0x6f3/0x8f0
	...

RIP  [<ffffffff8150dca2>] xfrm_input+0x3c2/0x5a0
Kernel panic - not syncing: Fatal exception in interrupt

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
---

	Patch is in final testing, I'm sending it opportunistically.
	Mike, please chim in if this patch didn't help in the end!

 net/xfrm/xfrm_input.c |   10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

--- a/net/xfrm/xfrm_input.c
+++ a/net/xfrm/xfrm_input.c
@@ -238,11 +238,6 @@ int xfrm_input(struct sk_buff *skb, int nexthdr, __be32 spi, int encap_type)
 
 		skb->sp->xvec[skb->sp->len++] = x;
 
-		if (xfrm_tunnel_check(skb, x, family)) {
-			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMODEERROR);
-			goto drop;
-		}
-
 		spin_lock(&x->lock);
 		if (unlikely(x->km.state == XFRM_STATE_ACQ)) {
 			XFRM_INC_STATS(net, LINUX_MIB_XFRMACQUIREERROR);
@@ -271,6 +266,11 @@ int xfrm_input(struct sk_buff *skb, int nexthdr, __be32 spi, int encap_type)
 
 		spin_unlock(&x->lock);
 
+		if (xfrm_tunnel_check(skb, x, family)) {
+			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMODEERROR);
+			goto drop;
+		}
+
 		seq_hi = htonl(xfrm_replay_seqhi(x, seq));
 
 		XFRM_SKB_CB(skb)->seq.input.low = seq;

Re: [PATCH] xfrm: fix xfrm_input/xfrm_tunnel_check oops

[Steffen Klassert]

On Thu, Apr 02, 2015 at 10:58:24AM +0300, Alexey Dobriyan wrote:
> https://bugzilla.kernel.org/show_bug.cgi?id=95211
> 
> Commit 70be6c91c86596ad2b60c73587880b47df170a41
> ("xfrm: Add xfrm_tunnel_skb_cb to the skb common buffer") added check
> which dereferences ->outer_mode too early but larval SAs don't have
> this pointer set (yet). So check for tunnel stuff later.
> 
> Mike Noordermeer reported this bug and patiently applied all the debugging.
> 
> Technically this is remote-oops-in-interrupt-context type of thing.
> 
> BUG: unable to handle kernel NULL pointer dereference at 0000000000000034
> IP: [<ffffffff8150dca2>] xfrm_input+0x3c2/0x5a0
> 	...
> [<ffffffff81500fc6>] ? xfrm4_esp_rcv+0x36/0x70
> [<ffffffff814acc9a>] ? ip_local_deliver_finish+0x9a/0x200
> [<ffffffff81471b83>] ? __netif_receive_skb_core+0x6f3/0x8f0
> 	...
> 
> RIP  [<ffffffff8150dca2>] xfrm_input+0x3c2/0x5a0
> Kernel panic - not syncing: Fatal exception in interrupt
> 
> Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>

Good catch!

Applied to the ipsec tree, thanks everyone!

Re: [PATCH] xfrm: fix xfrm_input/xfrm_tunnel_check oops

[Alexey Dobriyan]

On Tue, Apr 07, 2015 at 10:57:32AM +0200, Steffen Klassert wrote:
> On Thu, Apr 02, 2015 at 10:58:24AM +0300, Alexey Dobriyan wrote:
> > https://bugzilla.kernel.org/show_bug.cgi?id=95211
> > 
> > Commit 70be6c91c86596ad2b60c73587880b47df170a41
> > ("xfrm: Add xfrm_tunnel_skb_cb to the skb common buffer") added check
> > which dereferences ->outer_mode too early but larval SAs don't have
> > this pointer set (yet). So check for tunnel stuff later.
> > 
> > Mike Noordermeer reported this bug and patiently applied all the debugging.
> > 
> > Technically this is remote-oops-in-interrupt-context type of thing.
> > 
> > BUG: unable to handle kernel NULL pointer dereference at 0000000000000034
> > IP: [<ffffffff8150dca2>] xfrm_input+0x3c2/0x5a0
> > 	...
> > [<ffffffff81500fc6>] ? xfrm4_esp_rcv+0x36/0x70
> > [<ffffffff814acc9a>] ? ip_local_deliver_finish+0x9a/0x200
> > [<ffffffff81471b83>] ? __netif_receive_skb_core+0x6f3/0x8f0
> > 	...
> > 
> > RIP  [<ffffffff8150dca2>] xfrm_input+0x3c2/0x5a0
> > Kernel panic - not syncing: Fatal exception in interrupt
> > 
> > Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
> 
> Good catch!
> 
> Applied to the ipsec tree, thanks everyone!

For the record Mike confirmed that the patch works:
https://bugzilla.kernel.org/show_bug.cgi?id=95211#c20

Re: [PATCH] xfrm: fix xfrm_input/xfrm_tunnel_check oops

[Romain Francoise]

On Tue, Apr 07, 2015 at 02:43:41PM +0300, Alexey Dobriyan wrote:
> On Tue, Apr 07, 2015 at 10:57:32AM +0200, Steffen Klassert wrote:
>> On Thu, Apr 02, 2015 at 10:58:24AM +0300, Alexey Dobriyan wrote:
>> > https://bugzilla.kernel.org/show_bug.cgi?id=95211
>> > 
>> > Commit 70be6c91c86596ad2b60c73587880b47df170a41
>> > ("xfrm: Add xfrm_tunnel_skb_cb to the skb common buffer") added check
>> > which dereferences ->outer_mode too early but larval SAs don't have
>> > this pointer set (yet). So check for tunnel stuff later.
>> > 
>> > Mike Noordermeer reported this bug and patiently applied all the debugging.
>> > 
>> > Technically this is remote-oops-in-interrupt-context type of thing.
>> > 
>> > BUG: unable to handle kernel NULL pointer dereference at 0000000000000034
>> > IP: [<ffffffff8150dca2>] xfrm_input+0x3c2/0x5a0
>> > 	...
>> > [<ffffffff81500fc6>] ? xfrm4_esp_rcv+0x36/0x70
>> > [<ffffffff814acc9a>] ? ip_local_deliver_finish+0x9a/0x200
>> > [<ffffffff81471b83>] ? __netif_receive_skb_core+0x6f3/0x8f0
>> > 	...
>> > 
>> > RIP  [<ffffffff8150dca2>] xfrm_input+0x3c2/0x5a0
>> > Kernel panic - not syncing: Fatal exception in interrupt
>> > 
>> > Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
>> 
>> Good catch!
>> 
>> Applied to the ipsec tree, thanks everyone!

> For the record Mike confirmed that the patch works:
> https://bugzilla.kernel.org/show_bug.cgi?id=95211#c20

Does this only affect configurations that use VTI? It's not clear from
the description.

Thanks.

Re: [PATCH] xfrm: fix xfrm_input/xfrm_tunnel_check oops

[Steffen Klassert]

On Tue, Apr 07, 2015 at 02:14:43PM +0200, Romain Francoise wrote:
> On Tue, Apr 07, 2015 at 02:43:41PM +0300, Alexey Dobriyan wrote:
> > On Tue, Apr 07, 2015 at 10:57:32AM +0200, Steffen Klassert wrote:
> >> On Thu, Apr 02, 2015 at 10:58:24AM +0300, Alexey Dobriyan wrote:
> >> > https://bugzilla.kernel.org/show_bug.cgi?id=95211
> >> > 
> >> > Commit 70be6c91c86596ad2b60c73587880b47df170a41
> >> > ("xfrm: Add xfrm_tunnel_skb_cb to the skb common buffer") added check
> >> > which dereferences ->outer_mode too early but larval SAs don't have
> >> > this pointer set (yet). So check for tunnel stuff later.
> >> > 
> >> > Mike Noordermeer reported this bug and patiently applied all the debugging.
> >> > 
> >> > Technically this is remote-oops-in-interrupt-context type of thing.
> >> > 
> >> > BUG: unable to handle kernel NULL pointer dereference at 0000000000000034
> >> > IP: [<ffffffff8150dca2>] xfrm_input+0x3c2/0x5a0
> >> > 	...
> >> > [<ffffffff81500fc6>] ? xfrm4_esp_rcv+0x36/0x70
> >> > [<ffffffff814acc9a>] ? ip_local_deliver_finish+0x9a/0x200
> >> > [<ffffffff81471b83>] ? __netif_receive_skb_core+0x6f3/0x8f0
> >> > 	...
> >> > 
> >> > RIP  [<ffffffff8150dca2>] xfrm_input+0x3c2/0x5a0
> >> > Kernel panic - not syncing: Fatal exception in interrupt
> >> > 
> >> > Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
> >> 
> >> Good catch!
> >> 
> >> Applied to the ipsec tree, thanks everyone!
> 
> > For the record Mike confirmed that the patch works:
> > https://bugzilla.kernel.org/show_bug.cgi?id=95211#c20
> 
> Does this only affect configurations that use VTI? It's not clear from
> the description.

xfrm_tunnel_check() dereferences outer_mode only if it finds a tunnel
pointer, so it affects only vti SAs.