1 $Id: OSPF-ALIGNMENT.txt,v 1.1 2004/11/17 17:59:52 gdt Exp $
3 Greg Troxel <gdt@ir.bbn.com>
6 The OSPF specification (RFC2328) and the OSPF Opaque LSA specification
7 (RFC2370) are ambiguous about LSAs whose data section is not an
8 integral multiple of 4 octets. This note examines the issue and
9 proposes clarifications to ensure interoperability.
11 RFC2328 does not specify that LSA lengths be a multiple of 4.
12 It does not require that LSAs in update packets be aligned.
13 However, all structures defined by RFC2328 are multiples of 4, and
14 thus update packets with those structures must be aligned.
15 LSA length is defined in Appendix A.4 as
18 The length in bytes of the LSA. This includes the 20 byte LSA
21 RFC2370 defines Opaque LSAs, which are intended to contain arbitrary
24 This memo defines enhancements to the OSPF protocol to support a new
25 class of link-state advertisements (LSA) called Opaque LSAs. Opaque
26 LSAs provide a generalized mechanism to allow for the future
27 extensibility of OSPF. Opaque LSAs consist of a standard LSA header
28 followed by application-specific information. The information field
29 may be used directly by OSPF or by other applications. Standard OSPF
30 link-state database flooding mechanisms are used to distribute Opaque
31 LSAs to all or some limited portion of the OSPF topology.
36 Opaque LSAs contain some number of octets (of application-specific
37 data) padded to 32-bit alignment.
39 This can be interpreted in several ways:
41 A) The payload may be any number of octets, and the length field
42 reflects the payload length (e.g. length 23 for 3 octets of payload),
43 but there are padding octets following the LSA in packets, so that the
44 next LSA starts on a 4-octet boundary. (This approach is common in
45 the BSD user/kernel interface.)
47 B) The payload must be a multiple of 4 octets, so that the length is a
48 multiple of 4 octets. This corresponds to an implementation that
49 treats an Opaque LSA publish request that is not a multiple of 4
52 C) The payload can be any number of octets, but padding is added and
53 included in the length field. This interpretation corresponds to an
54 OSPF implementation that accepts a publish request for an Opaque LSA
55 that is not a multiple of 4 octets. This interpretation is
56 nonsensical, because it claims to represent arbitrary lengths, but
57 does not actually do so --- the receiver cannot distinguish padding
60 D) Accept according to A, and transmit according to B.
62 Option A arguably violates RFC 2328, which doesn't say anything about
63 adding padding (A.3.5 shows a diagram of adjacent LSAs which are shown
64 as all multiples of 4). This option is thus likely to lead to a lack
67 Option B restricts what data can be represented as an Opaque LSA, but
68 probably not in a serious way. It is likely to lead to
69 interoperability in that the complex case of non-multiple-of-4 lengths
72 However, an implementation that follows A and emits an LSA with
73 payload length not a multiple of 4 will not interoperate with an
74 Option B implementation.
76 Given that all known and documented uses of Opaque LSAs seem to be
77 multiples of 4 octets, we choose Option B as the clarification.
83 In section A.4, add a second sentence about length:
86 The length in bytes of the LSA. This includes the 20 byte LSA
87 header. The length must be an integral multiple of 4 bytes.
89 Add to the list in Section 13:
91 Verify that the length of the LSA is a multiple of 4 bytes. If
92 not, discard the entire Link State Update Packet.
98 Opaque LSAs contain some number of octets (of application-specific
99 data) padded to 32-bit alignment.
103 Opaque LSAs contain some a number of octets (of
104 application-specific data). The number of octets must be a
110 At BBN, we use Opaque LSAs to exchange data among routers; the format
111 of the data is naturally aligned to 4 bytes, and thus does not raise
112 this issue. We created a test program to publish Opaque data via IPC
113 to the OSPF daemon (quagga), and this program accepts strings on the
114 command line to publish. We then used this test program to publish
115 software version strings. Quagga's ospfd then crashed on a
116 NetBSD/sparc64 machine with an alignment fault, because the odd-length
117 LSAs were marshalled into a link-state update packet with no padding.
118 While this behavior was a clear violation of RFC2380, it was not clear
119 how to remedy the problem.