[Top] [All Lists]

Re: Allowing netlink_family to be any integer (was: [PATCH 2.6] iptables

To: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>
Subject: Re: Allowing netlink_family to be any integer (was: [PATCH 2.6] iptables CLUSTERIP target)
From: jamal <hadi@xxxxxxxxxx>
Date: 22 Oct 2004 08:53:02 -0400
Cc: Thomas Graf <tgraf@xxxxxxx>, Harald Welte <laforge@xxxxxxxxxxxxx>, lmb@xxxxxxx, ahu@xxxxxxx, hch@xxxxxxxxxxxxx, netdev@xxxxxxxxxxx, "David S. Miller" <davem@xxxxxxxxxxxxx>
In-reply-to: <20041022122533.GA9713@xxxxxxxxxxxxxxxxxxx>
Organization: jamalopolous
References: <20041021142527.GG3551@xxxxxxxxxxxxxxxxxxxxxxx> <E1CKkWZ-0005x5-00@xxxxxxxxxxxxxxxxxxxxxxxx> <20041021225315.GH19714@xxxxxxxxxxxxxx> <20041022122533.GA9713@xxxxxxxxxxxxxxxxxxx>
Reply-to: hadi@xxxxxxxxxx
Sender: netdev-bounce@xxxxxxxxxxx
On Fri, 2004-10-22 at 08:25, Herbert Xu wrote:
> On Fri, Oct 22, 2004 at 12:53:15AM +0200, Thomas Graf wrote:
> > 
> > The only real problem I see is sk_protocol being only 8bit as
> > limiting factor.
> Well we wouldn't be using sk_protocol to look things up anymore.
> We'd be holding a reference on the family directly once the socket
> is created.

Thats why i said:
This does not exclude the use of the netlink numbers, but should ease
them. I think that rtnetlink is already overloaded and would benefit
from having a clean split. As an example things like links and IPaddress
should probably reside elsewhere since they are generic enough a

Orthogal to this:
If we could get more people to use the work from Evgeniy as opposed to
creating a new netlink protocols, then it would benefit people - Harald
could have used that scheme in what he was trying to do for example.


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