tech-net archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: IP_HDRINCL byte ordering



Looks good in concept to me, with a few comments/questions.

  - Does this date from 4.4BSD?  Did FreeBSD have a flag day?

  - I would like to see in the IP_HDRINCL and IP_HDRINCL_RAW comments
    something like
      (struct ip in host order)
      (struct ip in network order)
    added.

  - The suggestion for portable should only include coping, not
    erroring.  "portable code that only works on the latest
    not-yet-released version" is an oxymoron in my book.

  - I would like to see the new #define and support code added, and then
    see tests be ok on all platforms for a bit, and then convert the
    in-tree users to the new codepoint.  Probably will amount to the
    same thing, but seems safer.

  - the traceroute code seems to turn on byteswap macros presumably
    because it varies by platform.  That's ok if I got it right.
    However, then we need to use IP_HDRINCL_RAW only, and not the old
    one.  This is in-tree code so presumably it should just hard-use the
    new define and thus fail to compile on old systems.

  - for mrouted, same.   If we are going to htons when making headers,
    we can't use the old define
    


Home | Main Index | Thread Index | Old Index