Clarification request for draft-brezak-win2k-krb-rc4-hmac-04: keyusage numbers and state chaining

Luke Howard lukeh at PADL.COM
Fri Nov 29 02:35:01 EST 2002


With regard to draft-brezak-win2k-krb-rc4-hmac-04.txt:

1. In Section 8.3 (GSSAPI WRAP semantics), "seq" is undefined:

	// new encryption key salted with seq
	Kcrypt = HMAC(Kcrypt, (int32)seq);

   Presumably this is the big-endian sequence number (Token.SEND_SEQ)
   but this should be clarified, particularly given that elsewhere
   little endian salts are used.

2. Should not the padding for exportable keys be nine, not
   seven, bytes for the effective key length to be 56 bits?

	memset(Kcrypt+7, 0xab, 9);

   vs.

	memset(Kcrypt+7, 0xab, 7);

   This applies to section 8.2 as well as 8.3.

3. What key derivation salt is used for the authorization data
   signatures in draft-brezak-win2k-krb-authz-00.txt? No mapping
   for KRB5_KU_AD_KDC_ISSUED / KRB5_KU_AUTH_DATA_TICKET_EXTENSION
   is provided.

-- Luke


--
Luke Howard | PADL Software Pty Ltd | www.padl.com



More information about the krbdev mailing list