netdev
[Top] [All Lists]

Re: [PATCH] IPV6: note on shared socket options

To: dlstevens@xxxxxxxxxx
Subject: Re: [PATCH] IPV6: note on shared socket options
From: YOSHIFUJI Hideaki / 吉藤英明 <yoshfuji@xxxxxxxxxxxxxx>
Date: Thu, 05 Feb 2004 11:09:55 +0900 (JST)
Cc: davem@xxxxxxxxxx, netdev@xxxxxxxxxxx
In-reply-to: <OF2F9C078B.064A8A19-ON87256E30.0077B005-88256E30.0077AF76@xxxxxxxxxx>
Organization: USAGI Project
References: <20040204.203739.16838230.yoshfuji@xxxxxxxxxxxxxx> <OF2F9C078B.064A8A19-ON87256E30.0077B005-88256E30.0077AF76@xxxxxxxxxx>
Sender: netdev-bounce@xxxxxxxxxxx
In article 
<OF2F9C078B.064A8A19-ON87256E30.0077B005-88256E30.0077AF76@xxxxxxxxxx> (at Wed, 
4 Feb 2004 13:49:51 -0800), David Stevens <dlstevens@xxxxxxxxxx> says:

>       These are defined by draft-ietf-magma-msf-api-03.txt to be in
> netinet/in.h (no place else). What's the advantage of adding another
> copy in in6.h (which currently isn't used by anything)? Portable user apps
> must include netinet/in.h for them, and in-kernel code already does.

(I assume you're talking about the "alternatives.")

Kernel do not use netinet/*.h.

My main point is, do not let people (or myself) forget reserved 
(or used) range.  So, it is enough for me to add a comment on that.


Well, I really do not think that the name of MCAST_xxx is good.
Numeric assignment of "MCAST_xxx" functionalities 
is only required to be unique in that level in theory
since we put MCAST_xxx at network level (IPPROTO_{IP,IPV6} for now).
However, they are share just because of the name.


Anyway, what I want to add is some small note on shared range.
Thanks.

-- 
Hideaki YOSHIFUJI @ USAGI Project <yoshfuji@xxxxxxxxxxxxxx>
GPG FP: 9022 65EB 1ECF 3AD1 0BDF  80D8 4807 F894 E062 0EEA

<Prev in Thread] Current Thread [Next in Thread>