| 1 |
On Mon, 2004-05-24 at 17:52, Aron Griffis wrote: |
| 2 |
> Chris Gianelloni wrote: [Mon May 24 2004, 05:08:08PM EDT] |
| 3 |
> > Is there a reason why we're only starting DHCP on one ethX device? On |
| 4 |
> > many systems, the FireWire is getting detected before the NIC, which is |
| 5 |
> > loading eth1394 as eth0. We either need FireWire loading last, or to |
| 6 |
> > set the system up to do DHCP on more than one interface. I *definitely* |
| 7 |
> > prefer the latter, as it is much more flexible. |
| 8 |
> |
| 9 |
> If we do something like that, then we should stop attempting once one |
| 10 |
> device gets an address. |
| 11 |
|
| 12 |
I'm just curious, but why? |
| 13 |
|
| 14 |
Allow me to give you two examples from my own home. |
| 15 |
|
| 16 |
1. My firewall - the NIC connected to my cable modem is eth1 even |
| 17 |
though I *have* an eth0 *and* an eth2, both of which are active. |
| 18 |
|
| 19 |
2. My AMD-MP box - the on-board NIC is 10/100 and I am using a gige |
| 20 |
NIC, so it is eth0 |
| 21 |
|
| 22 |
I still think it would be just as easy for us to try dhcp on all of the |
| 23 |
NICs detected. |
| 24 |
|
| 25 |
Also, plasmaroo has already given me a "test" of the next genkernel to |
| 26 |
try out, which removes eth1394 from loading before hotplug, so NICs |
| 27 |
should get eth0 before firewire. Couple that with a blacklist on |
| 28 |
eth1394, and the only way it gets brought up is manually. |
| 29 |
|
| 30 |
-- |
| 31 |
Chris Gianelloni |
| 32 |
Developer |
| 33 |
Games/LiveCD Teams |
| 34 |
Gentoo Linux |
| 35 |
|
| 36 |
Is your power animal a penguin? |
Archives