Re: [PATCH v2] net: ipv6: fix skb_over_panic in __ip6_append_data

[Willem de Bruijn <willemdebruijn.kernel@gmail.com>]

On Thu, Mar 10, 2022 at 5:30 PM Jakub Kicinski <kuba@kernel.org> wrote:
>
> On Thu, 10 Mar 2022 14:13:28 -0800 Tadeusz Struk wrote:
> > diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
> > index 4788f6b37053..6d45112322a0 100644
> > --- a/net/ipv6/ip6_output.c
> > +++ b/net/ipv6/ip6_output.c
> > @@ -1649,6 +1649,16 @@ static int __ip6_append_data(struct sock *sk,
> >                       skb->protocol = htons(ETH_P_IPV6);
> >                       skb->ip_summed = csummode;
> >                       skb->csum = 0;
> > +
> > +                     /*
> > +                      *      Check if there is still room for payload
> > +                      */
>
> TBH I think the check is self-explanatory. Not worth a banner comment,
> for sure.
>
> > +                     if (fragheaderlen >= mtu) {
> > +                             err = -EMSGSIZE;
> > +                             kfree_skb(skb);
> > +                             goto error;
> > +                     }
>
> Not sure if Willem prefers this placement, but seems like we can lift
> this check out of the loop, as soon as fragheaderlen and mtu are known.
>
> >                       /* reserve for fragmentation and ipsec header */
> >                       skb_reserve(skb, hh_len + sizeof(struct frag_hdr) +
> >                                   dst_exthdrlen);

Just updating this boundary check will do?

        if (mtu < fragheaderlen ||
            ((mtu - fragheaderlen) & ~7) + fragheaderlen <
sizeof(struct frag_hdr))
                goto emsgsize;