On 2020-10-11 11:31, Jakub Kicinski wrote:
On Sat, 10 Oct 2020 12:51:30 -0400 Willem de Bruijn wrote:
> > @@ -1385,7 +1386,17 @@ static int ibmveth_poll(struct napi_struct *napi,
int budget)
> > skb_put(skb, length);
> > skb->protocol = eth_type_trans(skb, netdev);
> >
> > - if (length > netdev->mtu + ETH_HLEN) {
> > + /* PHYP without PLSO support places a -1 in the ip
> > + * checksum for large send frames.
> > + */
> > + if (be16_to_cpu(skb->protocol) == ETH_P_IP) {
You can byteswap the constant, so its done at compilation time.
Thanks for the comments.
For V2 of patch I will change above to BE16_TO_CPU()
> > + struct iphdr *iph = (struct iphdr
*)skb->data;
> > +
> > + iph_check = iph->check;
>
> Check against truncated/bad packets.
.. unless I missed context. Other code in this driver seems to peek in
the network and transport layer headers without additional geometry
and integrity checks, too.
Good catch, even if we trust the hypervisor to only forward valid
frames this needs to be at least mentioned in the commit message.
Also please add Fixes tags to both patches.
For V2: ( posting soon )
-Will add Fix tags
-update commit message re: validity of frames from Hypervisor.
-switch be16_to_cpu() to BE16_TO_CPU().
Thanks
