Re: init scripts/md5sum
Bernhard Rosenkraenzer <[email protected]> Mon, 28 Mar 2005 00:54:57 +0200
| Newsgroups | gmane.linux.arklinux.devel |
|---|---|
| Organization | Ark Linux team |
| Message-ID | <[email protected]> |
On Sunday 27 March 2005 20:53, Luke wrote:
> [12:46:02] <AeroClown> if you look in modprobe.conf you will probably find
> an alias eth0 <driver>
> [12:46:09] <magpie> k
> [12:46:18] <AeroClown> if the init scripts where working right then your
> card would be working
> [12:46:31] <AeroClown> you can just do
> [12:46:34] <nick01> hi clown
> [12:46:38] <AeroClown> nano /etc/modprobe.conf
> [12:46:47] <AeroClown> look from the line that says alias eth0
> [12:46:52] <AeroClown> and the driver name should follow it
> [12:47:24] <AeroClown> they still aren't real sure as of yet as to why all
> of a sudden its not loading up the drivers correctly
At least on the 2 boxes where I could reproduce it:
the problem was that /etc/sysconfig/network-scripts/network-functions used
alias=`modprobe -c | awk "/^(alias|install) $1 / { print \\$3 }"`
to figure out which module to load.
Now on some systems modprobe.conf doesn't say alias[space]eth0[space]whatever,
but alias[tab]eth0[tab]whatever, and the awk expression wouldn't catch that.
I've changed the statement to
alias=`modprobe -c | awk "/^(alias|install)[[:space:]]+$1[[:space:]]+/{ print
\\$3 }"`
in initscripts-7.50-5ark, and I'm fairly sure ppl who are still seeing this
problem still have initscripts-7.50-4ark and lower.
I'll build a new iso with this change as soon as the new kernel is built,
there were a couple of things worth fixing in the old one too (most notably a
rockridge CD parsing error).
LLaP
bero