On Thu, Oct 8, 2020 at 12:04 PM Willem de Bruijn <willemdebruijn.ker...@gmail.com> wrote: > > On Thu, Oct 8, 2020 at 1:34 PM Cong Wang <xiyou.wangc...@gmail.com> wrote: > > > > The uninit data is allocated by packet_alloc_skb(), if dev->hard_header_len > > is 0 and 'len' is anything between [0, tunnel->hlen + sizeof(struct iphdr)), > > dev_validate_header() still returns true obviously but only 'len' > > bytes are copied > > from user-space by skb_copy_datagram_from_iter(). Therefore, those bytes > > within range (len, tunnel->hlen + sizeof(struct iphdr)] are uninitialized. > > With dev->hard_header_len of zero, packet_alloc_skb() only allocates len > bytes. > > With SOCK_RAW, the writer is expected to write the ip and gre header > and include these in the send len argument. The only difference I see > is that with hard_header_len the data starts reserve bytes before > skb_network_header, and an additional tail has been allocated that is > not used. > > But this also fixes a potentially more serious bug. With SOCK_DGRAM, > dev_hard_header/ipgre_header just assumes that there is enough room in > the packet to skb_push(skb, t->hlen + sizeof(*iph)). Which may be > false if this header length had not been reserved. > > Though I've mainly looked at packet_snd. Perhaps you are referring to > tpacket_snd?
I think what Cong means is that hard_header_len has to be set properly to prevent an AF_PACKET/RAW user from sending a frame that is too short (shorter than the header length). When an AF_PACKET/RAW user sends a frame shorter than the header length, and the code on the sending path still expects a full header, it will read uninitialized data. If my understanding is right, I agree on this part. However, there's something I don't understand in the GRE code. The ipgre_header function only creates an IP header (20 bytes) + a GRE base header (4 bytes), but pushes and returns "t->hlen + sizeof(*iph)". What is t->hlen? It seems to me it is the sum of t->tun_hlen and t->encap_hlen. What are these two?
