svn commit: r308184 - stable/11/sys/netinet

[Hiren Panchasara]

Author: hiren
Date: Tue Nov  1 22:40:25 2016
New Revision: 308184
URL: https://svnweb.freebsd.org/changeset/base/308184

Log:
  MFC r307545
  Make sure tcp_mss() has the same check as tcp_mss_update() to have t_maxseg set
  to at least 64.
  
  This is still just a coverup to avoid kernel panic and not an actual fix.
  
  PR:			213232
  Sponsored by:		Limelight Networks

Modified:
  stable/11/sys/netinet/tcp_input.c
Directory Properties:
  stable/11/   (props changed)

Modified: stable/11/sys/netinet/tcp_input.c
==============================================================================
--- stable/11/sys/netinet/tcp_input.c	Tue Nov  1 22:38:26 2016	(r308183)
+++ stable/11/sys/netinet/tcp_input.c	Tue Nov  1 22:40:25 2016	(r308184)
@@ -3754,7 +3754,15 @@ tcp_mss(struct tcpcb *tp, int offer)
 			(void)sbreserve_locked(&so->so_snd, bufsize, so, NULL);
 	}
 	SOCKBUF_UNLOCK(&so->so_snd);
-	tp->t_maxseg = mss;
+	/*
+	 * Sanity check: make sure that maxseg will be large
+	 * enough to allow some data on segments even if the
+	 * all the option space is used (40bytes).  Otherwise
+	 * funny things may happen in tcp_output.
+	 *
+	 * XXXGL: shouldn't we reserve space for IP/IPv6 options?
+	 */
+	tp->t_maxseg = max(mss, 64);
 
 	SOCKBUF_LOCK(&so->so_rcv);
 	if ((so->so_rcv.sb_hiwat == V_tcp_recvspace) && metrics.rmx_recvpipe)