On Sun, Nov 18, 2018 at 8:02 PM Herbert Xu <herb...@gondor.apana.org.au> wrote:
>
> On Fri, Nov 16, 2018 at 01:32:50PM -0800, Cong Wang wrote:
> >
> > This is true only when there is a skb_checksum_init*() or
> > skb_checksum_validate*() prior to it, it seems not true for
> > nf_ip_checksum() where skb->csum is correctly set to pesudo header
> > checksum but there is no validation of the original skb->csum.
> > So this check should be still inverted there??
> >
> > Or am I still missing anything here?
>
> What do you mean? My copy of nf_ip_checksum seems to be doing the
> right thing as far as verifying CHECKSUM_COMPLETED goes.
Hmm, it calls csum_tcpudp_magic() directly instead of
__skb_checksum_validate_complete(), but it also sets ip_summed
to CHECKSUM_UNNECESSARY:
20 if ((protocol == 0 && !csum_fold(skb->csum)) ||
21 !csum_tcpudp_magic(iph->saddr, iph->daddr,
22 skb->len - dataoff, protocol,
23 skb->csum)) {
24 skb->ip_summed = CHECKSUM_UNNECESSARY;
25 break;
26 }
which means the rx checksum fault won't be triggered.
