Permalink
Commits on Sep 23, 2003
  1. assume presense of getifaddrs

    itojun committed Sep 23, 2003
  2. bound check for sockaddr *

    itojun committed Sep 23, 2003
  3. simplify getifaddr loop

    itojun committed Sep 23, 2003
  4. sizeof(size_t) != sizeof(void *)

    itojun committed Sep 23, 2003
Commits on Sep 22, 2003
  1. mark persistent SP "persistent".

    XXX in theory refcnt should do the right thing, however, we have "spdflush"
    which would touch all SPs.  another solution would be to de-register persistent
    SPs from sptree.
    itojun committed Sep 22, 2003
Commits on Sep 21, 2003
  1. 2003-09-21 JINMEI, Tatuya <jinmei@isl.rdc.toshiba.co.jp>

    	* kame/sys/netinet6/ip6_output.c (ip6_ctloutput): simplified the
    	case of getsockopt IPV6_PKTOPTIONS by simply returning empty
    	data.  Note that the change does not make any difference from the
    	application's point of view, since we've already quitted storing
    	the options on a socket.
    
    2003-09-21  JINMEI, Tatuya  <jinmei@isl.rdc.toshiba.co.jp>
    	* kame/sys/netinet6/ip6_input.c (ip6_savecontrol): simplified the
    	function interface as a cleanup.  This change should not make any
    	difference from the API point of view.  Many other kernel files
    	that call this function were changed accordingly.
    jinmei committed Sep 21, 2003
  2. removed IPv6 inputopts.

    jinmei committed Sep 21, 2003
  3. removed inputopts.

    jinmei committed Sep 21, 2003
  4. removed inputopts from inpcb.

    jinmei committed Sep 21, 2003
  5. removed the in6p_inputopts member from in6pcb.

    jinmei committed Sep 21, 2003
  6. removed the inp6_inputopts member from inpcb.

    jinmei committed Sep 21, 2003
  7. simplified getsockopt(IPV6_PKTOPTIONS):

    * RFC3542 (effectively) deprecated the
    * semantics of the 2292-style pktoptions.
    * Since it was not reliable in nature (i.e.,
    * applications had to expect the lack of some
    * information after all), it would make sense
    * to simplify this part by always returning
    * empty data.
    jinmei committed Sep 21, 2003
  8. a 2292bis cleanup:

      now ip6_savecontrol() simply takes mbuf ** just like ip_savecontrol(), since
      RFC3542 doesn't require to keep history of control information.
      removed ip6_recvpktopts{} accordingly.
    jinmei committed Sep 21, 2003
  9. simplify by getifaddrs. rm unused function.

    itojun committed Sep 21, 2003
  10. fix compilation on netbsd/openbsd

    itojun committed Sep 21, 2003
  11. Compile without RANDOM_IP_ID option

    t-momose committed Sep 21, 2003
Commits on Sep 20, 2003
  1. 2003-09-20 JINMEI, Tatuya <jinmei@isl.rdc.toshiba.co.jp>

    	* kame/kame/rtadvd/if.c (if_nametosdl): made sure to free the
    	temporary buffer.
    	In response to KAME PR 502 reported by Takayuki Sakuma.
    jinmei committed Sep 20, 2003
  2. made sure to free the temporary buffer in if_nametosdl()

    reported by takayuki@finet.fujitsu.com.
    jinmei committed Sep 20, 2003
Commits on Sep 19, 2003
  1. 2003-09-19 JINMEI, Tatuya <jinmei@isl.rdc.toshiba.co.jp>

    	* kame/sys/netinet6/ipsec.c (ipsec_copypkt): avoided coping m_flag
    	after M_MOVE_PKTHDR().  It would clear M_PKTHDR from the copy
    	target and cause unexpected behavior.  It is highly recommended
    	for IPsec ESP or IPcomp users to apply this fix.
    jinmei committed Sep 19, 2003
  2. do not copy m_flag after M_MOVE_PKTHDR(). it would clear M_PKTHDR fro…

    …m the
    
    copy target.
    (this part should have been removed with M_COPY_PKTHDR.)
    jinmei committed Sep 19, 2003
  3. realloc pedant

    itojun committed Sep 19, 2003
Commits on Sep 17, 2003
  1. 2003/09/17 23:37:38 JST kjc@csl.sony.co.jp

            * eliminate reserved qids (class handles).
              allow cbq, hfsc and priq to use arbitrary non-zero qids.
            * make the qid assignment similar to the pf_tag assignment.
              this is a step to share the logic with pf_tag.
    kjc committed Sep 17, 2003
  2. reenable ALTQ

    kjc committed Sep 17, 2003
  3. make the qid assignment similar to the pf_tag assignment.

     - temporarily use tagname functions for qnames until we have a common
       framework for both tagname and qname.
    move pf_altq_running from altq_subr.c to pf_ioctl.c.
     - swap the order in the altq ioctl case statement so that the altq commit
       case can fall into the altq start case.
    kjc committed Sep 17, 2003
  4. the reserved qids (class handles) are gone.

    kjc committed Sep 17, 2003
  5. eliminate reserved qids (class handles).

    allow cbq, hfsc and priq to use arbitrary non-zero qids.
    this is a step for making the qid assignment similar to
    the pf_tag assignment.
    kjc committed Sep 17, 2003
  6. freebsd4 has ip_randomid()

    itojun committed Sep 17, 2003
  7. freebsd[45] has ip_randomid()

    itojun committed Sep 17, 2003
  8. ip_id.c conflicts

    itojun committed Sep 17, 2003
  9. set conflict to ip_id.c

    itojun committed Sep 17, 2003
  10. use ip_randomid() for netbsd

    itojun committed Sep 17, 2003
Commits on Sep 16, 2003
  1. aes-xcbc-mac is now RFC3566

    itojun committed Sep 16, 2003
  2. xcbc-mac is now RFC3566

    itojun committed Sep 16, 2003