Messages in this thread | | | Date | Mon, 27 Mar 2000 21:53:38 -0700 | From | Richard Gooch <> | Subject | Re: PCI probing in slot order? |
| |
David L. Nicol writes: > Richard Gooch wrote: > > > > David L. Nicol writes: > > > > > > you mean, the way *BSD does it? > > > > Dunno. How does *BSD do it? > > when installing freeBSD, you need to identify the hardware you are > using. The different kinds of network cards have different > abbreviations for purposes of configuring them, instead of eth0 > being my network card regardless of the particular card. The ne2000 > is ne0, the smc is sm0 (or something) and so on. To ease > configuration for startuup files and so forth, there is a big > settings file where you edit it (or a tool edits it for you) to set > $ethdevice_0=ne0 or something like that and then all the scripts > refer to, like,
Erk! Yuk! How painful. Well, I'm definately not suggesting that. If we were to move net interfaces into the normal device namespace, there would be a /dev/net directory which would just have "eth0", "eth1", "ppp0" and so on. That will provide the easy interface (for the typical case where you only have one interface).
> OSF1 is that way too, if you swap your hardware you have to > adjust the configuration, not just the kernel modules.
Blech!
> I like the kernel taking care of that abstraction, it makes creating > a single configuration I can put on heterogenous hardware easier.
Indeed. We don't want to lose that. But it would be nice to be also be able to reference interfaces based on their location. I've set up two routers (each with 3 interfaces) where both "logical" guesses were wrong (the order was in fact eth0, eth2, eth1 going from left to right).
Regards,
Richard.... Permanent: rgooch@atnf.csiro.au Current: rgooch@ras.ucalgary.ca
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |