In the course of struggling with this problem, I have created a new custom kernel. This was already happening when I tried using the GUI tool Linux Kernel Configurator, which is a kind of graphical front end for editing a .Config file much like 'make xconfig'. I finally figured out what it did and did some studying (LDP Documentation and Linux Programming Unleashed again) on compiling a custom kernel, which is something I wanted to do anyway. In a sense, the module I wanted to make adustments to is part of the kernel too, so I guess there was no avoiding it. It can also be compiled into the kernel statically.

So I compiled this new kernel. The kernel compilation was configured so that my adjusted module was to be added as a module. The source code for the module that I made small adustments to compiled with no problems and can be loaded with 'insmod' with no problem. And, I can make further adjustments to the source code and use 'make modules' and 'make_install' to apply the changes.

This part of the problem is SOLVED.

Also, the Hello World modules are loading just fine now. The mysterious 'version 2.4.18-5custom' that shows up in the errors is the aborted kernel that was started when I used the Linux Kernel Configurator. The Makefile in the /build directory adds 'custom' to the version name for the kernel to keep the new compilation from overwriting the old kernel.

So, this part of the problem is also SOLVED.

I am also having some other odd version problems elsewhere, but that can be for another thread. And, shouldn't I be able to compile this module without making a new kernel? But, unless anyone else has anything to add,

PROBLEM SOLVED

Thanks to Mara and all who tuned in to help.

You should be able to compile the module for any kernel (2.4.x in your case) when you have source of this kernel.

I need to add a few points to this particular problem, because the issue has come up again. I am able to compile the module that was the original source of this inquiry, but only in the 2.4.18-5custom kernel. This was satisfactory, because it was mostly experimental. But now, I am back in my main 2.4.18-5 kernel and I am trying to compile and install another module and I am getting the same "unresolved symbols" errors.

In a way, this is a good thing, because it forces the issue back to the kernel and maybe some headers. The module I was first trying to compile was not the problem.

So what is going on? Something that comes to mind is that when I use my Red Hat Upgrade Agent, I am finding that my kernel upgrade is to 2.4.18 with no '-', or build number, or whatever they call that number. Has my kernel library been damaged or misplaced or something?