Please help: Odd mouse problem with kernel-2.6.7-1.494.2.2

[William W. Austin]

I have some mixed eide/scsi boxen which all repeatably and uniformly  
display the same odd symptom.  I use a ps/2 scroll mouse on all (5) of  
my systems which are pretty nearly identical (2 at work, 3 at home  
office), and since installing the 2.6.7-1.494.2.2 kernel the scroll  
wheel on the mouse is somewhat hosed.

Immediately after a hard reboot, it works normally; however after 5-15  
minutes of 'normal' work (xterms, email, web browsing, etc.), the  
system stops recognizing the scroll wheel altogether, and it never  
"mysteriously starts recognizing" it again thereafter (well not for up  
to 18 hours so far, anyway).

This is fully repeatable on all 5 boxen, and seems to be independent of  
the mouse in use (have tried 4 different ps/2 scroll-wheel mice).  It  
did not occur on any previous kernel or under fc1 or rh9.  It is also  
independent of whether gnome or kde desktop is used, and it is  
independent of which x-server is used ("nv" or the commercial/non-open- 
source "nvidia" driver from nvidia.com).  Finally, it is fully  
independent of whether I am connecting the mouse directly to the  
computer or from the kvm switch to the computer. (3 machines on one kvm  
switch, 2 on the other)

While at first glance it sounds highly illogical that this could be a  
kernel-related issue (sounds more like an x-server issue to me), I'm  
running out of other alternatives - I have both 2.6.5-1.358 (the  
original distribution one) and 2.6.7-1.494.2.2 kerels on these boxes:  
and if I boot the former it never occurs while if I boot the latter, it  
invariably occurs.

The messages in /var/log are essentially identical regardless of which  
kernel is in use, and no error messages show up in anything in /var/log  
which I can find.

I have bugzilla'ed this one, but any suggestions or other info anyone  
has on this one would be greatly appreciated - I need to run the later  
kernel to pick up some other fixes, and the scrollwheel stoppage is a  
royal pain.

(BTW, I have no idea whether it occurred in any of the update kernels  
between the two listed above: due to a byte-swap problem in the  
intervening kernels {bugzilla 126391}, none of them would allow my  
system to boot.)

-- 
William W. Austin 			   waustin at speakeasy.net
"Life is just a phase I'm going through... this time, anyway..."