[lkml]   [1998]   [Feb]   [2]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    SubjectRe: DHCPD v1.3 needs testing (fwd)
    In article <> you wrote:
    : >Dhcpcd must generate ethernet frames itself

    : And Token-Ring frames, and whatever frametypes you might
    : imagine using DHCP with ? I'm sorry, but I don't understand
    : why a DHCP client needs to know about all sorts of underlying
    : network frames,

    (AF_PACKET, SOCK_DGRAM) does NOT depend on link layer details.

    Really, it was supposed that client should open two sockets
    (AF_PACKET, SOCK_DGRAM) and bootp udp.
    When an address is acquired and checked for validity,
    it is assigned to interface, packet socket is unbound and
    background communication goes via udp.

    : when there is a perfectly capable network
    : stack that already has all of the necessary code.

    This stack has no code to work in IP promiscuous mode.

    2.0 had some bits of it, but it had "small" :-) drawback.
    It killed all the IP, when at least one interface
    was not configured.

    Now we have a chance to write true dhcpd/bootpd,
    which is able to work on multihomed host or router.

    Alexey Kuznetsov

     \ /
      Last update: 2005-03-22 13:41    [W:0.017 / U:16.012 seconds]
    ©2003-2017 Jasper Spaans. hosted at Digital OceanAdvertise on this site