15 posts and still not sorted :-(.
I had one of these a long time back on older hardware. It was with a network card. I eventually found windows selected irq11 and linux sgave it irq12, making it invisible.My box had aa via chipset and was giving the same 2 halfassed irqs to everything

Try booting with noapic on the command line. sometimes it needed
acpi=off noapic

I must have been running X for the working lsmod, because I just did it again for both working and non-working boots, and the only difference was in the order of some of the entries. When I ran it with X, those two modules showed up at the top.

I find it interesting that the mouse, which always works, seem to show up in lsmod as psmouse, but there is never any mention of the keyboard at all.

I tried modprobe usbhid, but it complained of an "invalid argument". usbkbd was not found.

business_kid: I guess I'll try that, but I'm not sure why you suggest it. I'm doing some reading about those options, but could you explain why you think that would help?

OK on the i810 module - I am sure that you were just in X, but that leaves us with no extra or missing modules when it fails or works.

I will try to look more into the usbhid module - it may require some parameters, but I was hoping it would just load and work.

Like you, I am stil confused with what module, if any, is handling the keyboard input - nothing in the lsmod list appears related. So I kind of assume it is being handled by the kernel and the BIOS directly - but that is a little outside my current knowledge base. But the point is, if we can identify the driver - or force it to use one, then we will have a handle to grab it by.

I can't speak for business_kid, but what the APIC and ACPI handle are interrupts and power control, respectively. I have had to disable ACPI on a few systems, usually to enable ethernet devices which can quit working otherwise. It is a shot in the dark - but it often hits a target!

The acpi allows several parameters, but usually adding 'acpi=off' to your boot append parameters will help if anything does. If your USB device is PCI the 'pci=noacpi' might also help.

To disable APIC add 'noapic' to the boot parameters.

For a full list of kernel params if you have the kernel source installed you will probably find it in:

/usr/src/linux/Documentation/kernel-parameters.txt

I'll look for a better understanding of the usbhid module - my thinking being that if we get the keyboard under 'our' own control we can figure out what is happening.

You might want to try the APIC and ACPI boot parameters, too (and read the docs). I have found the ACPI options to be the most helpful for non-working hardware, but of curse, YMMV!

And one last question - is that a Dell dude? In my readings I found mention of a particular Dell keyboard that had no work around - and would intermittently fail on boot... any chance of that?

It may help or change things. With the APIC off, (APIC is the Interrupt Controller) it's back to 16 interrupts. Most hardware problems of this sort are interrupt related. Now you are in the position where mouse and keyboard have to share the same interrupt on a ps2 port. It's a change in interrupts you can painlessly make. Have you a via chipset by any chance?
On early Via chipsets (Up to 2002 at least) the via chipset had a poor APIC which gave the same 2 halfassed irqs to everything. Try 'cat /proc/interrupts'

Nope, it doesn't seem to have an effect. It booted with those params and still had no keyboard. Anyway, I like having the power button play nicely with the OS. I had to restart it through the network.

I'm glad to hear it. It does probably eliminate interrupts as the issue.Now you say windows works these things. Can you find what resources windows uses for them? They might work in linux also.

According to the Device Manager, the keyboard uses the drivers kbdclass.sys and kbdhid.sys, and the mouse uses mouclass.sys and mouhid.sys, all apparently provided by Microsoft. I just tried it on my Vista machine, which has never seen this setup, and it worked fine there, even before it finished its little "installing device drivers" routine. So I guess its standard stuff.

It's the resources tab I would have been interested in. Sounds like you have vista. How sad a version of windows that is. I have it myself.

Try issuing the following as a lilo option at boot time:-

If it works, then add it to lilo.conf

I've noticed the same problems as the op, but I'm running -current64 on a (fairly new Toshiba) laptop. Every once in a while, the keyboard just does not work but it does in LiLO. Will be following this thread.

@business_kid: I have XP on the computer in question. I've just done most of my posting from Vista, on the other side of the desk. I don't seem to have a resources tab in XP, either, at least not in the driver properties dialog.

@vdemuth: I tried that and had several problem-free boots, but then the boot crashed with a kernel panic error, roughly:Keyboard and mouse did nothing, nor did the power button. I had to pull the plug. It booted okay after that (disk check went fine), but I took those options out of lilo.conf (put there for convenience while I repeatedly restarted). I guess it was worth a shot.

@manwichmakesameal: I knew there had to be others!

*bump*