[net] netfilter: xt_TCPMSS: handle CHECKSUM_COMPLETE in tcpmss_tg6()

From: Eric Dumazet <edumazet@google.com>

In case MSS option is added in TCP options, skb length increases by 4.
IPv6 needs to update skb->csum if skb has CHECKSUM_COMPLETE,
otherwise kernel complains loudly in netdev_rx_csum_fault() with a
stack dump.

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 net/netfilter/xt_TCPMSS.c |    9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

On Fri, Jan 15, 2016 at 8:21 AM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> From: Eric Dumazet <edumazet@google.com>
>
> In case MSS option is added in TCP options, skb length increases by 4.
> IPv6 needs to update skb->csum if skb has CHECKSUM_COMPLETE,
> otherwise kernel complains loudly in netdev_rx_csum_fault() with a
> stack dump.
>
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> ---
>  net/netfilter/xt_TCPMSS.c |    9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
>
> diff --git a/net/netfilter/xt_TCPMSS.c b/net/netfilter/xt_TCPMSS.c
> index b7c43def0dc6..e118397254af 100644
> --- a/net/netfilter/xt_TCPMSS.c
> +++ b/net/netfilter/xt_TCPMSS.c
> @@ -228,7 +228,7 @@ tcpmss_tg6(struct sk_buff *skb, const struct xt_action_param *par)
>  {
>         struct ipv6hdr *ipv6h = ipv6_hdr(skb);
>         u8 nexthdr;
> -       __be16 frag_off;
> +       __be16 frag_off, oldlen, newlen;
>         int tcphoff;
>         int ret;
>
> @@ -244,7 +244,12 @@ tcpmss_tg6(struct sk_buff *skb, const struct xt_action_param *par)
>                 return NF_DROP;
>         if (ret > 0) {
>                 ipv6h = ipv6_hdr(skb);
> -               ipv6h->payload_len = htons(ntohs(ipv6h->payload_len) + ret);
> +               oldlen = ipv6h->payload_len;
> +               newlen = htons(ntohs(oldlen) + ret);
> +               if (skb->ip_summed == CHECKSUM_COMPLETE)
> +                       skb->csum = csum_add(csum_sub(skb->csum, oldlen),
> +                                            newlen);
> +               ipv6h->payload_len = newlen;

Probably should have a utility function do this. Maybe something like
update_hdr_field(skb, old, new)? Could return new value so above could
just be:

ipv6h->payload_len = update_hdr_field(skb, ipv6h->payload_len,
htons(ntohs(oldlen) + ret);

>         }
>         return XT_CONTINUE;
>  }
>
>

On Fri, 2016-01-15 at 08:43 -0800, Tom Herbert wrote:

> Probably should have a utility function do this. Maybe something like
> update_hdr_field(skb, old, new)? Could return new value so above could
> just be:
> 
> ipv6h->payload_len = update_hdr_field(skb, ipv6h->payload_len,
> htons(ntohs(oldlen) + ret);

Well, oldlen would still need to be ipv6h->payload_len

Not sure a helper will be nice for a stable submission.

Probably better for net-next ?

On Fri, Jan 15, 2016 at 9:18 AM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> On Fri, 2016-01-15 at 08:43 -0800, Tom Herbert wrote:
>
>> Probably should have a utility function do this. Maybe something like
>> update_hdr_field(skb, old, new)? Could return new value so above could
>> just be:
>>
>> ipv6h->payload_len = update_hdr_field(skb, ipv6h->payload_len,
>> htons(ntohs(oldlen) + ret);
>
> Well, oldlen would still need to be ipv6h->payload_len
>
That comes from the second argument.

> Not sure a helper will be nice for a stable submission.
>
> Probably better for net-next ?
>
I think this would apply to ipv6: update skb->csum when CE mark is
propagated also, maybe others? Seems safer to implement the logic only
in one place if possible... No big deal though.

>

From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Fri, 15 Jan 2016 09:18:44 -0800

> On Fri, 2016-01-15 at 08:43 -0800, Tom Herbert wrote:
> 
>> Probably should have a utility function do this. Maybe something like
>> update_hdr_field(skb, old, new)? Could return new value so above could
>> just be:
>> 
>> ipv6h->payload_len = update_hdr_field(skb, ipv6h->payload_len,
>> htons(ntohs(oldlen) + ret);
> 
> Well, oldlen would still need to be ipv6h->payload_len
> 
> Not sure a helper will be nice for a stable submission.
> 
> Probably better for net-next ?

Agreed.  We already have those csum_replace{2,4}() helpers in net/checksum.h
which we could augment with ones that can handle length changes too.

On Fri, Jan 15, 2016 at 02:04:49PM -0500, David Miller wrote:
> From: Eric Dumazet <eric.dumazet@gmail.com>
> Date: Fri, 15 Jan 2016 09:18:44 -0800
> 
> > On Fri, 2016-01-15 at 08:43 -0800, Tom Herbert wrote:
> > 
> >> Probably should have a utility function do this. Maybe something like
> >> update_hdr_field(skb, old, new)? Could return new value so above could
> >> just be:
> >> 
> >> ipv6h->payload_len = update_hdr_field(skb, ipv6h->payload_len,
> >> htons(ntohs(oldlen) + ret);
> > 
> > Well, oldlen would still need to be ipv6h->payload_len
> > 
> > Not sure a helper will be nice for a stable submission.
> > 
> > Probably better for net-next ?
> 
> Agreed.  We already have those csum_replace{2,4}() helpers in net/checksum.h
> which we could augment with ones that can handle length changes too.

I'm going to place this into the nf tree, thanks!