It seems this is an xorg-server 1.4 issue:
This happens to me using a microsoft internet keyboard and i noticed this interesting post on Gentoo forums as i play with gentoo a bit.
Hope it helps to know if it happens to you too and don't blame -current

edit: reading a bit more about the bug i found:
source: bugs.gentoo

I'm not sure if it's ok to post links those urls, just trying to show references and hope it helps.

edit2:
@Robby: not fixed on xorg-server-1.4-i486-2
another report: https://bugs.freedesktop.org/show_bug.cgi?id=12434

mannyslack:
Is this fixed in the new build that came out today? I was under the impression that it would be, although I don't see any indication of it in the sources. It's a known bug, either way.

in this bug report, https://bugs.freedesktop.org/show_bug.cgi?id=12434
someone has managed to fix the issue, but it cause the autorepeater didn't work as expected

Thanks for the links, guys - I'll forward them to Pat.
For whatever reason, I can't seem to access the freedesktop site at the moment, so I don't know what the content of the link is. Assuming it's not an official patch, I'm guessing that Pat will wait for one before fixing this in -current.

a new branch of 1.4 will comes up soon, so i guess PV should wait for this branch rather than using the trivial patch

Upgraded my test system to current. Worked through the issues with pixman. I'm left with two issues--

The first is that I can't get kernel-generic to boot. The root filesystem is ext2. Using the command for mkinitrd that works with the 21.5 kernel ginves an initrd.gz file significantly larger than previously.
The boot dies in the process of using initrd.gs. The console shows
---------------
---------------
The system then stops in a limited shell with no access to the drives.

Booting without initrd.gs ends with
---------------
---------------

No problems with kernel-huge.

Second, kde is looking for some library files that I can't locate. Haven't had any problems with the things I've tried to do to this point.

I have just finished upgrading to Current :-)

Two small extra settings in rc.modules were needed:

1. /sbin/modprobe psmouse proto=imps
2. /sbin/modprobe eepro100

Most things seem to work nicely. So now I'll test my IBM T23 deeper during next days.

No, you haven't. See below.

See /etc/modprobe.d/psmouse -- the psmouse module is no longer blacklisted in the default /etc/modprobe.d/blacklist, and the load options are in the /etc/modprobe.d/psmouse file. However, if you have both the old blacklist file and the blacklist.new file present, since modprobe(8) looks at *every* file in /etc/modprobe.d/ for module load options, the psmouse module is still blacklisted.

Does the e100 module not work for your hardware? - it's the preferred one now, as eepro100 is now blacklisted.

Vise versa.

Just thought it would be nice to share my experience upgrading to slackware-current - listing the problems I encountered and how I solved them.

Warning: This is not meant to encourage you to change to -current. If you need a stable system (desktop or server), keep with the stable version of slackware. Changing to -current is for the adventurous who don't mind (and do find it fun, like me) running into problems and then trying to fix them.

1) I used slackroll to upgrade the packages (just following standard procedures)
2) changed the default runlevel in inittab to 3 (knew there were probably going to be some problems with X)
3) rebooted the first time using the new 'huge' kernel
4) corrected lilo.conf and ran mkinitrd for the new 'generic' kernel


Up to here, everything ok - system booting new kernel
Tried to start X - no success.

Discovered that I forgot about pixman (and I had read the post in this thread!). Installed it, but still no X.
Changed my driver to "vesa" in xorg.conf -> x started ok.

Recompiled my openchrome driver, now X complained that there was an incompatibility between the openchrome driver and the xserver (major ABI versions different).
Started x with "startx -- -ignoreABI" and it worked!

Put the same 'ignoreABI' option in /etc/kde/kdm/kdmrc:
ServerArgsLocal=-nolisten tcp -ignoreABI"

Other small adjustments:
- had to get vmware-any-any-update114 and rerun the runme.pl script to recompile some things (just hit enter to all questions)
- needed to add the normal users to the audio group for sound to work (after starting X a window popped up complaining about accessing a device)

Result: running slackware-current now and it seems to by very stable!

I deleted obsolete historical backups of blacklist files from
/etc/modprobe.d that earlier caused me to use extra settings in rc.modules. I wasn't fully aware of the rules of this place.

Option for mouse (options psmouse proto=imps) is now "commented" in /etc/modprobe.d/psmouse file and module eepro100 is not blacklisted in /etc/modprobe.d/blacklist.

All of the above gives me 100% current stable system as seen of today

Thanks

Yep. Just damn. Next time maybe I'll actually look at the files involved instead of going from memory. So close, yet so far away... ;-)
Anyway, the original post is fixed.

Good!

You say the psmouse line is commented -- why? I thought you wanted the psmouse module to load with the imps protocol? Note that there is a new /etc/rc.d/rc/modules shipped with the kernel-modules package too; it no longer forceloads the psmouse module.

Anyway, sorry about the accidental reversal of the e100 and eepro100 modules in my original post above - however, I'm curious - does the e100 module not work with your hardware?

My Logitech mouse (along with T23 trackpoint) is working well. All mouse support section in my rc.modules is commented out (disabled). In /etc/modprobe.d/psmouse file everything is also commented out. What about our IMPS protocol ? I don't know...

Invoking: $ lsmod | grep mouse
gives me:
psmouse 40720 0



I tried both e100 and eepro100 modules by modprobing and rmmod-ing them and both seem to work. I checked it only by issuing # ifconfig that shows usable network interfaces.

Any news on the status of XOrg bug 12434 - keyboard LED's don't reflect modifier state?
Is Pat going to wait for a point release of xorg-server, or going to use the patch?