Proposed Change - RFC - net-tools to iproute2

James Robertson jwrober at
Tue Jun 29 20:16:54 PDT 2004

Matthew Burgess wrote:

> On Sat, 26 Jun 2004 16:05:57 -0700
> "Jim Gifford" <lfs at> wrote:
>>I'm currently proposing a change, I have made this change on my system
>>and it does effect performance.
>>I propose the move from net-tools to iproute2.
> I'd like to see this go in a.s.a.p.  I've looked over the thread and
> there seem to be to be 3 main points against immediate inclusion:
> 1) tools provided by net-tools which don't have a replacement in
> iproute2
> * I've had a look at and `netstat` `route` and
> `ifconfig` would all appear to be covered by the `ip` binary, at least
> the functionality I'd expect in a *base* system, i.e. LFS, is described
> in  However, that
> said, I'm am not by any means an authority when it comes to networking
> issues, so feel free to point out that this isn't in fact adequate for a
> base system.  My opinion is based on: 1) Can I bring an interface
> up/down 2) Can I see which interfaces are up 3) Can I view & change the
> routing tables.  I'd be interested to hear from you if there are any
> core networking features I've left out here that a base system should be
> able to handle.

I read all the thread - I have a question.  If the `ip` binary can do 
`netstat` `route` and `ifconfig`, why can't we put instructions in for 
symbolic links for these to `ip`?  Good educational value IMO.

> 2) Dependency on Berkeley DB
> * If we suppress the building of `arp` then we no longer require db. 
> I'd assume that ARP isn't too common a requirement for a base system,
> and that you'll know if you need it.  As long as we point to BLFS (and
> BLFS has instructions for rebuilding iproute2 with db support) then I
> don't see this as an issue.

I agree with the rest of the thread.  If you are doing hard-core 
networking support or something, then you will need `arp` and know that 
you need it.  I am not a fan of putting db in the LFS book for one 
little tool/utility.



More information about the lfs-dev mailing list