svn commit: r340267 - stable/12/sys/netinet

[Glen Barber]

Author: gjb
Date: Thu Nov  8 21:58:23 2018
New Revision: 340267
URL: https://svnweb.freebsd.org/changeset/base/340267

Log:
  MFC r340260 (emaste):
   Avoid buffer underwrite in icmp_error
  
   icmp_error allocates either an mbuf (with pkthdr) or a cluster depending
   on the size of data to be quoted in the ICMP reply, but the calculation
   failed to account for the additional padding that m_align may apply.
  
   Include the ip header in the size passed to m_align.  On 64-bit archs
   this will have the net effect of moving everything 4 bytes later in the
   mbuf or cluster.  This will result in slightly pessimal alignment for
   the ICMP data copy.
  
   Also add an assertion that we do not move m_data before the beginning of
   the mbuf or cluster.
  
  Approved by:	re (kib)
  Security:	CVE-2018-17156
  Sponsored by:	The FreeBSD Foundation

Modified:
  stable/12/sys/netinet/ip_icmp.c
Directory Properties:
  stable/12/   (props changed)

Modified: stable/12/sys/netinet/ip_icmp.c
==============================================================================
--- stable/12/sys/netinet/ip_icmp.c	Thu Nov  8 21:56:06 2018	(r340266)
+++ stable/12/sys/netinet/ip_icmp.c	Thu Nov  8 21:58:23 2018	(r340267)
@@ -313,7 +313,8 @@ stdreply:	icmpelen = max(8, min(V_icmp_quotelen, ntohs
 #endif
 	icmplen = min(icmplen, M_TRAILINGSPACE(m) -
 	    sizeof(struct ip) - ICMP_MINLEN);
-	m_align(m, ICMP_MINLEN + icmplen);
+	m_align(m, sizeof(struct ip) + ICMP_MINLEN + icmplen);
+	m->m_data += sizeof(struct ip);
 	m->m_len = ICMP_MINLEN + icmplen;
 
 	/* XXX MRT  make the outgoing packet use the same FIB
@@ -355,6 +356,8 @@ stdreply:	icmpelen = max(8, min(V_icmp_quotelen, ntohs
 	 * reply should bypass as well.
 	 */
 	m->m_flags |= n->m_flags & M_SKIP_FIREWALL;
+	KASSERT(M_LEADINGSPACE(m) >= sizeof(struct ip),
+	    ("insufficient space for ip header"));
 	m->m_data -= sizeof(struct ip);
 	m->m_len += sizeof(struct ip);
 	m->m_pkthdr.len = m->m_len;