>>>>> On Thu, 17 Jul 2003 14:21:41 -0400,
>>>>> Vladislav Yasevich <[EMAIL PROTECTED]> said:
>>> The reason is that
>>> most Mobile IPv6 implementation are supported in the "kernel" and
>>> already do checksumming.
>>
>> What exactly do you imagine about the implementation that supports
>> mobile IPv6 in the kernel?
>> If it generates, does checksumming, and
>> sends a packet with a mobile header completely within the kernel (the
>> KAME implementation apparently acts like this), it can do so without
>> conflicting with any API spec. So I don't get why this is the reason
>> for specifying the behavior on a raw socket used by applications
>> (i.e., not by the kernel.)
> I guess I was originally looking at it from the recieve side. On
> a aystem that already implements MIPv6 message handling at the
> network layer (in the kernel), it's rather pointless to set the
> IPV6_CHECKSUM option since the packet will be checked anyway.
I see, but my point still applies: if the implementation receives,
validates the checksum, and processes a packet with a mobile header
completely within the kernel, it can do so without conflicting with
any API spec.
The KAME implementation apparently acts like this. It in fact does
not send the received packet with a mobility header to the "raw IPv6
socket" layer. Instead, it directly passes the packet to a separate
routine dedicated to Mobile IPv6 where the checksum is validated.
JINMEI, Tatuya
Communication Platform Lab.
Corporate R&D Center, Toshiba Corp.
[EMAIL PROTECTED]
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
