Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
Commits on Nov 28, 2004
  1. corrected nroff syntax

    jinmei authored
  2. s/dhcp6sctl/dhcp6ctl/

    jinmei authored
  3. 2004-11-28 JINMEI, Tatuya <jinmei@isl.rdc.toshiba.co.jp>

    jinmei authored
    	* kame/kame/dhcp6: catch up draft-ietf-dhc-lifetime-02.
    	Specifically,
    	- renamed "lifetime" to "refreshtime".  Note that configuration
    	  compatibility was lost.
    	- implemented the default and minimum timer values
    	- clarified that the refresh time option can only be available for
    	  Information-request and reply exchanges
  4. new argument to dhcp6_set_options

    jinmei authored
  5. scoping-arch-02

    jinmei authored
  6. updated rev. number of rfc2462bis to 06

    jinmei authored
  7. use the default value for IRT when unspecified as per dhc-lifetime-02.

    jinmei authored
    also clarifies that IRT can only be used in response to info-req.
  8. do not specify the IRT in ORO except for information-request

    jinmei authored
    as specified in dhc-lifetime-02.
  9. regenerated

    jinmei authored
  10. s/LIFETIME/REFRESHTIME/

    jinmei authored
  11. s/lifetime/refresh time/ as per dhc-lifetime-02.

    jinmei authored
    note: backward compatibility to old configurations is not provided.
  12. removed overinitialization

    jinmei authored
  13. some cleanups

    jinmei authored
  14. 2004-11-28 JINMEI, Tatuya <jinmei@isl.rdc.toshiba.co.jp>

    jinmei authored
    	* kame/kame/libinet6/getaddrinfo.c (res_queryN): set an
    	appropriate limit of query timeouts in order to mitigate hopeless
    	delay due to misbehaving DNS servers ignoring AAAA queries.
    	getaddrinfo() now sends A queries first with this change, but the
    	query ordering does not affect the ordering of the resulting
    	addrinfo list.
    	(FreeBSD only at the moment)
  15. declare res_send_timeout()

    jinmei authored
  16. Set an appropriate limit of query timeouts for getaddrinfo(). We let

    jinmei authored
    the resolver routine decide the timeout policy until we get the first
    positive response.  After that, we explicitly set the maximum limit of
    timeouts to the larger one between 1 second and twice the round trip
    time of the previous successful query.  This way, we can mitigate
    hopeless delay due to misbehaving DNS servers that ignore AAAA queries
    (or queries of "unknown" types in general).
    
    (Note: FreeBSD only at the moment, since we also need to modify res_send.c)
Commits on Nov 27, 2004
  1. grammar

    jinmei authored
  2. 2004-11-27 JINMEI, Tatuya <jinmei@isl.rdc.toshiba.co.jp>

    jinmei authored
    	* kame/kame/libinet6/getaddrinfo.c (set_source): fixed a bug that
    	deprecated source addresses was not correctly identified for
    	address ordering (RFC3484).
  3. 2004-11-27 JINMEI, Tatuya <jinmei@isl.rdc.toshiba.co.jp>

    jinmei authored
    	* kame/kame/libinet6/getaddrinfo.c: suppress AAAA queries if the
    	node does not have IPv6 addresses other than the loopback and
    	link-local addresses.  Additionally, AI_ADDRCONFIG is now more
    	specification compliant in that getaddrinfo() only issues AAAA
    	queries when the node has non-loopback addresses.
  4. - suppress AAAA queries if the node does not have IPv6 addresses othe…

    jinmei authored
    …r than
    
      the loopback and link-local addresses.
    - the behavior with AI_ADDRCONFIG now more conforms to RFC3493 (while it is
      probably not really helpful).
    
    TODO:
      - man page update
      - configuration knob in resolv.conf?
Commits on Nov 26, 2004
  1. fixed a grammatical error

    suz authored
Commits on Nov 25, 2004
  1. stop linking getaddrinfo.c to use libc's getaddrinfo()

    suz authored
    (this is a temporary workaround to fix an getaddrinfo() failure)
Commits on Nov 22, 2004
  1. wording

    jinmei authored
Commits on Nov 19, 2004
  1. removed the following note from the change log on Nov18 by suz.

    jinmei authored
      (RFC2460 section 4.4 needs an update)
  2. removed comments about the "conflict" between a sanity check and the

    jinmei authored
    standard algorithm of processing a routing header.
Commits on Nov 18, 2004
  1. 2004-11-18 JINMEI, Tatuya <jinmei@isl.rdc.toshiba.co.jp>

    jinmei authored
    	* kame/sys/netinet6/in6.c (in6_if2idlen): introduced a new
    	function, which was previously a local module "if2idlen", and
    	called it when necessary.  With this change we do not store the
    	IFID length in in6_ifextra{}.  It did not always work as intended
    	due to initialization timing issues.
Something went wrong with that request. Please try again.