svn commit: r341258 - stable/10/sys/netinet

[Ed Maste]

Author: emaste
Date: Thu Nov 29 20:14:09 2018
New Revision: 341258
URL: https://svnweb.freebsd.org/changeset/base/341258

Log:
  MFC r340260: 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.
  
  Reported by:	A reddit user
  Security:	CVE-2018-17156
  Sponsored by:	The FreeBSD Foundation

Modified:
  stable/10/sys/netinet/ip_icmp.c
Directory Properties:
  stable/10/   (props changed)

Modified: stable/10/sys/netinet/ip_icmp.c
==============================================================================
--- stable/10/sys/netinet/ip_icmp.c	Thu Nov 29 19:28:01 2018	(r341257)
+++ stable/10/sys/netinet/ip_icmp.c	Thu Nov 29 20:14:09 2018	(r341258)
@@ -299,7 +299,8 @@ stdreply:	icmpelen = max(8, min(V_icmp_quotelen, ntohs
 	mac_netinet_icmp_reply(n, m);
 #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
@@ -341,6 +342,8 @@ 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;