Looking for some pointers, this is a follow on from a previous post, but thought it best to have its own thread.

I'm looking on information on how modules are inserted/loaded at boot time, what determines which ones are selected etc etc.

I know I have /lib/modules/[kernel version]
When I build my kernel, does the /lib/modules/[kernel version] filder beomce poulated based on the .config file used to customise the kernel?

I also have /etc/rc.d/rc.modules
What creates this file, is it standard or customised to my system?

I see /etc/rc.d/rc.S runs the above file at boot time, so what determines which modules are available and which should be loaded?

cat /proc/modules only shows:
agpgart 42648 0 (unused)



All help/comments are as always appreciated.

Yes, the modules directory is created when you launch I don't know how it is created, but if you have a look at this file you will see that it only should be used if udev is not present or not working correctly (or your hardware needs a special treatment).

It is udev that loads the modules.

My understanding of rc.modules is that it is hand-crafted. I don't know whether its contents are slackware specific or something generic that was originally from upstream somewhere, but the source file can be found in $mirror/slackware/source/k/kernel-modules/rc.modules.new. It hasn't been updated in a long while though, and when Patrick decided to modularise the loop device module he added an etc/rc.d/rc.loop initscript rather than use the traditional rc.modules file. Whether this means rc.modules is 'deprecated' or not I don't know.

I should add that these questions arose from findings on a slack 9.1 system 2.4.22 kernel.

Thought I'd add that as I read somewhere that udev was a > 2.4.22 kernel thing.

What decides which modules should be loaded at boot time?

Back during slackware 9.1, hotplug was the predecessor to udev. So if you aren't using rc.modules, or statically linking your modules, then look there.

FYI rc.modules is contained within the kernel-modules package.