Re: [PATCH] ipv6: check return value of ipv6_skip_exthdr

[Kees Cook <keescook@chromium.org>]

On Wed, Nov 17, 2021 at 07:16:10PM +0100, Jordy Zomer wrote:
> The offset value is used in pointer math on skb->data.
> Since ipv6_skip_exthdr may return -1 the pointer to uh and th
> may not point to the actual udp and tcp headers and potentially
> overwrite other stuff. This is why I think this should be checked.
> 
> Signed-off-by: Jordy Zomer <jordy@pwning.systems>
> ---
>  net/ipv6/esp6.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/net/ipv6/esp6.c b/net/ipv6/esp6.c
> index ed2f061b8768..dc4251655df9 100644
> --- a/net/ipv6/esp6.c
> +++ b/net/ipv6/esp6.c
> @@ -808,6 +808,11 @@ int esp6_input_done2(struct sk_buff *skb, int err)
>  		struct tcphdr *th;
>  
>  		offset = ipv6_skip_exthdr(skb, offset, &nexthdr, &frag_off);
> +
> +		if (offset < 0)
> +			err = -EINVAL;
> +			goto out;
> +

Ew. Yeah, it seems like ipv6_skip_exthdr() needs to be checked in a lot
of places. If this is part of protocol decoding, I'm surprised fuzzers
haven't found this? Is this state reachable?

I assume so, as there have been similar fixes in the past:
https://git.kernel.org/linus/92c6058024e87087cf1b99b0389d67c0a886360e

>  		uh = (void *)(skb->data + offset);
>  		th = (void *)(skb->data + offset);
>  		hdr_len += offset;
> -- 
> 2.27.0
> 

-- 
Kees Cook