Human interface devices

Keyboard (updated 24 September 2003)

The keyboard worked without trouble. I had the impression that it was connected internally via USB and would give me problems, but I was wrong and it's a PS/2 keyboard. There wasn't any harm in including USB keyboard modules.

The blue ``function keys'' on the laptop worked after I included ``Dell laptop support'' in the kernel (by setting CONFIG_I8K). I could also have included the module i8k, had I known about it. All of the function keys that map to keys on a vanilla PC keyboard (SysRq, Pause, NumLock and the numeric keypad) seem to produce the correct keycodes. The CD-eject key (Fn-F10) ejects the CD-ROM tray, and the screen brightness keys (Fn-up and Fn-down) adjust the screen brightness. Sometimes I have to press the CD-eject more than once, and sometimes changing the screen brightness ejects the CDROM. How strange.

I have just noticed (24 September 2003) that the ``magic SysRq'' shortcuts in the kernel are useless because the numeric keypad is accessed with the Fn key: since the Fn key is down to get SysRq, the Unmount, kIll, saK, showMem commands just change the loglevel. What a pain.

The only keys that don't work are the software volume controls, the suspend key, and the ``home page button'' above the keyboard. These pass keycodes to the operating system which you can see with showkey(1). Presumably you could program them if you were so inclined. This might be a good idea if you hit them accidentally and passing uncofigured keycodes does strange things to your system. Someone has documented how to do this but I have lost it.

Actually I'm not sure what the CRT/LCD key (Fn-F8) does. Connecting a CRT monitor to the port on the back sends simultaneous signals to the CRT and the LCD, irrespective of the status of that key. It doesn't produce a keycode, so the kernel is recognizing it, but I don't know what it does.

Keyboard: update 19 July 2004

I have learned that the Fn-key combinations are really read in by the BIOS. After upgrading to BIOS A30, as described in Section , the SysRq problem is solved by Fn-F10 no longer opens the CD Drive. Go figure.

Trackpad

The trackpad worked out of the box. It thinks it is a PS/2 mouse with two buttons. I am running gpm on the console and letting X watch the mouse via /dev/gpmdata. For some reason, gpm is on the second Debian CD, which meant I had to work extra hard to install it. I think it should be on the first so that one-disc installs can have mouse support on the console, but I don't make the CDs.

I have read that there is a driver by Symantec, who manufactures the pad, but I didn't need it.

Video Card: Needed More Recent Driver

This was the most difficult thing to get going. The video card is an ATI Radeon Mobility, but of recent enough vintage that the X server on the Debian 3.0 CD doesn't support it. I grabbed the xserver-xfree86 package from testing, which required me to also use the testing versions of libc6, libc6-dev, libdb1-compat, and locales. The X server now works using the ``radeon'' driver.

I used an apt ``Package Pin'' in /etc/apt/preferences to (hopefully) keep my testing packages up to date without running a full testing system that would require frequent updating. The pin is

Package: xserver-xfree86 libc6* libdb1-compat locales
Pin: release a=testing
Pin-Priority: 996

giving these packages a higher priority than the stable version at 995. I'm not sure if this will work automagically. We'll see.

For completeness, I'm running xserver-xfree86 version 4.2.1-6.