Re: make-dpkg: flavors and external modules
At 07:09 PM 03/26/02 +0100, martin f krafft wrote:
>> 2) When installing the package, is there a way to say, just add this to the
>> current list of kernels, but don't rotate the current kernel to .old?
>
>man kernel-img.conf
Thanks.
>but i usually just undo the change manually, it's quite easy...
Ok, I agree -- and then I'm sure what's happening. Most of the time I'm sure.
>> 3) what I don't understand is how I can boot different kernels, yet have
>> one /etc/modules and /etc/modules.con setup. Can someone explain?
>
>depmod -a is run at every boot.
I though depmod just when through my list of modules to find the
dependencies -- so that if I say load module modprobe foo it would know to
load module bar first, if foo depended on bar.
Let's say my I have a kernel without any modules -- everything compiled in.
If I then move to a modular kernel with, say, my NIC driver as a module
how does the new kernel know to load that module without me telling it?
Don't I need to run modconf first?
And going the other way, if I move from a module "m" to compiled in "y",
then I assume if my old /etc/modules says to load that module it will see
that it's already compiled in somehow?
I posted a week or so for a pointer that explains /etc/modules and
/etc/modules.conf and how it all works, but without response. Something
that steps me though what happens at boot.
>> 4) I'm using the nvida video driver. I'm sure this is the Wrong Way, but
>> currently after I create a new kernel I manually copy this driver into my
>> /lib/modules/*/kernel/video directory. Is there a trick to managing this
>> automatically so I don't have to do this manual copy?
>
>where's the driver from? doesn't the kernel provide an nvidia module
>itself?
No, it's only provided by Nvidia. I just followed the instructions at the
Nvidia site:
http://www.nvidia.com/view.asp?PAGE=linux
http://205.158.109.140/XFree86_40/1.0-2802/README
Which is:
Instructions for the Impatient:
$ tar xvzf NVIDIA_kernel.tar.gz
$ tar xvzf NVIDIA_GLX.tar.gz
$ cd NVIDIA_kernel
$ make install
$ cd ../NVIDIA_GLX
$ make install
>why not tell us which ones and i'll let you know if you need to employ
>initrd to enable modules being used before / is mounted.
So really, as long as I'm building in support for mounting my drive I don't
need to worry about it. initrd, if I understand, is just a temporary RAM
disk to use as a file system until to place the modules that needed to be
loaded to access my real disk. Mostly used for distributions that need to
have a general purpose kernel to deal with many types of hardware. That
about right?
My current module list is not very interesting, I was thinking more about
the *next* kernel I build ;)
:/lib/modules/2.4.17/kernel$ find .
.
./video
./video/NVdriver
./drivers
./drivers/scsi
./drivers/scsi/ide-scsi.o
./drivers/net
./drivers/net/8139too.o
./drivers/net/dummy.o
./drivers/sound
./drivers/sound/ac97_codec.o
./drivers/sound/sound.o
./drivers/sound/emu10k1
./drivers/sound/emu10k1/emu10k1.o
./drivers/usb
./drivers/usb/printer.o
./net
./net/ipv4
./net/ipv4/netfilter
./net/ipv4/netfilter/iptable_filter.o
./net/ipv4/netfilter/ip_tables.o
./net/ipv4/netfilter/ipt_multiport.o
Thanks for your time!
--
Bill Moseley
mailto:moseley@hank.org
--
To UNSUBSCRIBE, email to debian-user-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: