netdev
[Top] [All Lists]

Re: frag id byte order (PATCH)

To: Chris Wedgwood <cw@xxxxxxxx>
Subject: Re: frag id byte order (PATCH)
From: "David S. Miller" <davem@xxxxxxxxxx>
Date: Wed, 18 Apr 2001 12:23:40 -0700 (PDT)
Cc: Imran.Patel@xxxxxxxxx, netdev@xxxxxxxxxxx
In-reply-to: <20010418211440.A3894@metastasis.f00f.org>
References: <2D6CADE9B0C6D411A27500508BB3CBD0013C7923@eseis15nok> <15068.43560.329884.817170@pizda.ninka.net> <20010418211440.A3894@metastasis.f00f.org>
Sender: owner-netdev@xxxxxxxxxxx
Chris Wedgwood writes:
 > On Tue, Apr 17, 2001 at 01:40:08PM -0700, David S. Miller wrote:
 > 
 >     Imran.Patel@xxxxxxxxx writes:
 >      > Anycase, i have attached the patch (against ip6_output.c/kernel 
 > 2.4.3) below
 >      > (notice that some people will do anything to get their name in the 
 > kernel
 >      > sources:)
 >     
 >     I've applied your patch, thanks a lot.
 > 
 > I thought frag_id was an opaque value and hence so long as it is
 > consistent it need matter here?

It's opaque, sure, but not to things like TCP header compression et
al. :-)  (TCP header compression expects the network byte order value
of the IP header ID field to increase monotonically)

Besides, I'd rather all Linux machines spit out IDs in the same way
regardless of endianness.

Later,
David S. Miller
davem@xxxxxxxxxx

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