Here is a modified version of net.eth0 toat cna bring up adsl interfaces using adsl-start. This is a preliminatry version without any support for multiple adsl interfaces, hopefully when i get some time I will modify adsl-setup and this script to keep multiple profiles for different adsl interfaces. This needs the iface_IFACE set to "adsl" and a iface_user set to the username on the ISP in /etc/conf.d/net
Created attachment 14812 [details] the net.eth0 modified for adsl interfaces
Created attachment 14813 [details] the net.eth0 modified for adsl interfaces
need to add a check to make sure adsl-setup has been run as well, but this is a start.
Nay. net.eth? should not be responsible for ADSL connections. Perhaps pppoe.eth? will be better. You see, I actually use my eth0 for TWO purposes, one for PPPoE, the other for normal networking.
He modified it to also support ADSL besides the normal NICs, and NOT just to support ADSL only ...
Ok, so for what kind of DSL connections is this ? I just tried it with mine (eth0 is the NIC that is connected to my DSL Modem via pppoe) and it didn't work (I'm not complaining, just curious). AFAIK ppp0 is the standard interface for DSL rp-pppoe, am I correct ? So the DSL stuff might be more logical in net.ppp0.
You misunderstand. In my setup, net.adsl brings up ppp0 via PPPoE on eth0, and after that, net.eth0 brings up eth0. I know I shouldn't be doing this, but I don't have two network cables running to my server, only one. Interesting idea, but I agree that net.eth0 is the wrong place to put it. net.ppp0 or adsl.eth0 or pppoe.eth0 or net.pppoe0 are probably more appropriate.
Hi, I've been helping Martin maintain the rc-scripts stuff lately. I have a few comments about this bug. First, thanks for the patch to make adsl stuff work in net.eth0. We really appreciate patches. :-) That said, I have some questions. I have an ADSL connection personally, but I have a modem/router that does PPPoA and then does NAT for the rest of the network. In my setup there is no real interaction required from my workstation to bring up the ADSL connection. Therefore I don't quite understand what adsl-start and adsl-stop do. Would you mind explaining or pointing me to some documentation? Additionally some people in this bug have suggested that a better home would be net.ppp0 or that another script would be more appropriate. What was your rationale for putting this in net.eth0? Do you agree or disagree with their arguments? Sorry to put the burden back on you for explanation, but I can't apply this (or a variant) without a better understanding of it. Since this bug is so old, I'm closing NEEDINFO but please feel free to re-open if you can provide the needed information. Thanks!
PPPoE is different ... it requires things to be run in order to work on my routers which use PPPoE, i've made a small rc-script that runs adsl-start and adsl-stop ... these scripts are provided by the rp-pppoe package ... the only real package out there that handles PPPoE for you and does it nicely :) basically a typical setup looks like this: eth0 -> PPPoE -> ppp0 -> WAN eth1 -> LAN so when you run adsl-start, it sets up ppp0 for you ... eth0 is never brought up because thats just not how it works :) the ppp0 interface is run over eth0 so bringing up eth0 would kill the whole thing ... you dont need to add all the fancy stuff that we have for net.eth and net.ppp because all the settings are placed into /etc/ppp/pppoe.conf ... and all the times i've done PPPoE, network information was provided via dhcp so i guess the reason the patch is for net.eth0 is that eth0 is really used for bringing up the ppp0 and thus the router but in reality, `ifconfig eth0 up` does not happen ... the proposed stuff looks pretty good ... i would just remove the hardcoding of /etc/ppp/pppoe.conf from it since adsl-{start,stop} default to that anyways
Thanks for the explanation Spanky. I'll work on getting this in
What package provides adsl-start? It's not on my system
rp-pppoe
ok, I've done half of what needs to be done here: pidfile enforcement in the rp-pppoe package. That makes it possible to have separate pid files for different interfaces, and the pidfiles are controlled by net.eth0. This is important because net.eth0 needs to be able to shut down an interface even if the configuration has changed. In order for that to happen, there needs to be a reliable method for testing the existing configuration. I have the net.eth0 modifications ready as well, but they're on my laptop, so I'll commit them tomorrow to rc-scripts cvs
On this note, I've just created a bug detailing an idea I've been throwing around with regard to kernel PPPoE - check out bug #53954 if you're interested.
ok, after I made those changes to rp-pppoe I found a way to avoid the need for them in net.eth0. So eventually I'll back them out, once I'm certain about it. I've committed the necessary changes to net.eth0 and it will appear in the next release of baselayout. Thanks for the initial cut at the code, Patrick, and for your explanations, Spanky.
baselayout-1.10 is now in portage. It's package.masked at the moment but will be unmasked shortly.
