socket buf speedup from thorpej@netbsd, okay art@ ericj@:

Make insertion of data into socket buffers O(C):
* Keep pointers to the first and last mbufs of the last record in the
  socket buffer.
* Use the sb_lastrecord pointer in the sbappend*() family of functions
  to avoid traversing the packet chain to find the last record.
* Add a new sbappend_stream() function for stream protocols which
  guarantee that there will never be more than one record in the
  socket buffer.  This function uses the sb_mbtail pointer to perform
  the data insertion.  Make TCP use sbappend_stream(). On a profiling
run, this makes sbappend of a TCP transmission using
a 1M socket buffer go from 50% of the time to .02% of the time. Thanks
to Bill Sommerfeld and YAMAMOTO Takashi for their debugging
assistance!

1 files changed, 3 insertions, 3 deletions

diff --git a/sys/netinet/tcp_usrreq.c b/sys/netinet/tcp_usrreq.c
index 48704cb5a35..c18409f60bf 100644
--- a/sys/netinet/tcp_usrreq.c
+++ b/sys/netinet/tcp_usrreq.c
@@ -1,4 +1,4 @@
-/*	$OpenBSD: tcp_usrreq.c,v 1.63 2002/06/09 16:26:11 itojun Exp $	*/
+/*	$OpenBSD: tcp_usrreq.c,v 1.64 2002/08/08 17:07:32 provos Exp $	*/
 /*	$NetBSD: tcp_usrreq.c,v 1.20 1996/02/13 23:44:16 christos Exp $	*/
 
 /*
@@ -409,7 +409,7 @@ tcp_usrreq(so, req, m, nam, control)
 	 * marker if URG set.  Possibly send more data.
 	 */
 	case PRU_SEND:
-		sbappend(&so->so_snd, m);
+		sbappendstream(&so->so_snd, m);
 		error = tcp_output(tp);
 		break;
 
@@ -457,7 +457,7 @@ tcp_usrreq(so, req, m, nam, control)
 		 * of data past the urgent section.
 		 * Otherwise, snd_up should be one lower.
 		 */
-		sbappend(&so->so_snd, m);
+		sbappendstream(&so->so_snd, m);
 		tp->snd_up = tp->snd_una + so->so_snd.sb_cc;
 		tp->t_force = 1;
 		error = tcp_output(tp);