[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Scroll Wheel not working in v4... [Solved...sort of]



On Sun, 2005-07-03 at 09:25 +0200, Clorden wrote:
> On Sunday 03 July 2005 03:57, Tom Sightler wrote:
> Sorry I did not read this thread before. This KVM/mouse issue is an old 2.6.x 
> kernel related... ahem... "feature".
> 
> All 2.6.x kernels have problems when dealing with many KVM due to the fact 
> that certain KVMs do not fully respect the PS/2 standard while 2.6.x kernels 
> are much more "standard-compliant" than 2.4.x kernels.

I'm familiar with the "issue" of non-compliant KVM's.  That being said
these KVM's worked fine with 2.4 and don't with 2.6.  That's because
with 2.4 the PS/2 port driver simply forwarded data on to user space
while the 2.6 kernels input layer attempts to probe and detect the
device.

> The problem was solved in 2.6.11/12 kernels. If you are using a pre-2.6.11 
> kernel (such as the AS4's 2.6.9-11) you can avoid the mouse going crazy 
> booting with this parameter:
> 
> psmouse.proto=bare

Uh, I'm pretty sure this code certainly exists in the 2.6.9 kernel in
RHEL4.  Here are the lines from psmouse-base.c for 2.6.9-11

module_param_named(proto, psmouse_proto, charp, 0);
MODULE_PARM_DESC(proto, "Highest protocol extension to probe (bare,
imps, exps). Useful for KVM switches.");

The thing is, that's not really Ed's problem.  Ed didn't describe any
issue with his mouse "going crazy", only that his mouse wheel didn't
work.  Based on the output files he sent me, that's because the kernel
is detecting only a standard mouse when connected via the KVM.

My point is that the kernel deals with all types of buggy hardware.
Obviously these KVM's can support wheels since they work with the 2.4
kernel (and in my friend case Windows works fine) only 2.6 kernels don't
work.  I simple modification to the psmouse-base.c file to change the
module parameter to "Protocol to use" instead of "Highest Protocol to
Probe" would probably work around the problem.

If I did the patch I'd submit it to the upstream kernel and as a RFE for
RHEL4.  They might not take it, but there's a chance.

Later,
Tom



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]