On 20/04/15(Mon) 18:37, Mike Belopuhov wrote:
> On Tue, Apr 14, 2015 at 22:08 +0300, Lauri Tirkkonen wrote:
> > On Tue, Apr 14 2015 20:40:58 +0200, Mike Belopuhov wrote:
> > > According to 3.2 in RFC 7323:
> > >
> > > Once TSopt has been successfully negotiated, that is both <SYN> and
> > > <SYN,ACK> contain TSopt, the TSopt MUST be sent in every non-<RST>
> > > segment for the duration of the connection, and SHOULD be sent in an
> > > <RST> segment (see Section 5.2 for details). The TCP SHOULD remember
> > > this state by setting a flag, referred to as Snd.TS.OK, to one. If a
> > > non-<RST> segment is received without a TSopt, a TCP SHOULD silently
> > > drop the segment. A TCP MUST NOT abort a TCP connection because any
> > > segment lacks an expected TSopt.
> >
> > Thank you, I somehow missed the existence of this RFC.
> >
>
> Does anyone else want to comment on this?
Diff looks good to me. Since you mentioned other *BSD in your previous
post, I look at what Linux and Solaris do and they both seem to set
timestamps on keep alive packets. As for MD5 signatures Linux also
include it.
ok mpi@
> > > I had a stab at adding timestamp support to tcp_respond but couldn't
> > > test yet. If you feel like giving it a try, please be my guest.
> >
> > With your patch, I confirm that timestamps are present on keep-alive
> > messages.
> >
>
> The patch needs a small (but crucial) amendment: tcp pcb can be NULL...
>
> diff --git sys/netinet/tcp_subr.c sys/netinet/tcp_subr.c
> index c8c8e77..6f17af0 100644
> --- sys/netinet/tcp_subr.c
> +++ sys/netinet/tcp_subr.c
> @@ -295,13 +295,10 @@ tcp_template(tp)
> * attached mbufs.
> *
> * In any case the ack and sequence number of the transmitted
> * segment are as specified by the parameters.
> */
> -#ifdef INET6
> -/* This function looks hairy, because it was so IPv4-dependent. */
> -#endif /* INET6 */
> void
> tcp_respond(struct tcpcb *tp, caddr_t template, struct tcphdr *th0,
> tcp_seq ack, tcp_seq seq, int flags, u_int rtableid)
> {
> int tlen;
> @@ -370,14 +367,10 @@ tcp_respond(struct tcpcb *tp, caddr_t template, struct
> tcphdr *th0,
> xchg(th->th_dport, th->th_sport, u_int16_t);
> else
> flags = TH_ACK;
> #undef xchg
>
> - m->m_len = tlen;
> - m->m_pkthdr.len = tlen;
> - m->m_pkthdr.rcvif = (struct ifnet *) 0;
> th->th_ack = htonl(ack);
> th->th_x2 = 0;
> th->th_off = sizeof (struct tcphdr) >> 2;
> th->th_flags = flags;
> @@ -386,10 +379,26 @@ tcp_respond(struct tcpcb *tp, caddr_t template, struct
> tcphdr *th0,
> if (win > TCP_MAXWIN)
> win = TCP_MAXWIN;
> th->th_win = htons((u_int16_t)win);
> th->th_urp = 0;
>
> + if (tp && (tp->t_flags & (TF_REQ_TSTMP|TF_NOOPT)) == TF_REQ_TSTMP &&
> + (flags & TH_RST) == 0 && (tp->t_flags & TF_RCVD_TSTMP)) {
> + u_int32_t *lp = (u_int32_t *)(th + 1);
> + /* Form timestamp option as shown in appendix A of RFC 1323. */
> + *lp++ = htonl(TCPOPT_TSTAMP_HDR);
> + *lp++ = htonl(tcp_now + tp->ts_modulate);
> + *lp = htonl(tp->ts_recent);
> + tlen += TCPOLEN_TSTAMP_APPA;
> + th->th_off = (sizeof(struct tcphdr) + TCPOLEN_TSTAMP_APPA) >> 2;
> + }
> +
> + m->m_len = tlen;
> + m->m_pkthdr.len = tlen;
> + m->m_pkthdr.rcvif = (struct ifnet *) 0;
> + m->m_pkthdr.csum_flags |= M_TCP_CSUM_OUT;
> +
> /* force routing table */
> if (tp)
> m->m_pkthdr.ph_rtableid = tp->t_inpcb->inp_rtableid;
> else
> m->m_pkthdr.ph_rtableid = rtableid;
>