Laptop users with Synaptics touchpad, using kernel 2.6.x

Joshua Legbandt jtlegbandt at
Sun Aug 24 15:48:20 UTC 2003

On Sun, 2003-08-24 at 00:27, Mike A. Harris wrote:
> I've noticed a fair number of people lately who have been testing 
> the 2.6.x kernel on laptops, reporting they are unable to use 
> their laptop's touchpad anymore.
> This is a known issue which is related to ACPI.  If you have one 
> of these touchpads, and it does not work (most or all of them 
> wont work), then currently you have to download a separate GPL 
> licensed "synaptics" driver from the following website:
> In order to use these devices under 2.6.x kernels right now, you 
> MUST download this driver binary, or download the source and 
> compile it yourself.  I recommend using the binary driver as 
> compiling the source will be overly complex for most people.

Unfortunately, I can't seem to get the binary driver to work under
XFree86-4.3.0-22, which I am otherwise happy with (it's faster that the
RHL 9 and Severn packages by a good margin). The binary driver says that
it is compiled for 4.2.0 and looks like it loads from my logs, but
'"Protocol" "auto-dev"' doesn't work. From my limited debugging skills
(IANAP), it looks like this protocol should be provided by the driver.
I'll have to test with ACPI turned off to be sure that this isn't ACPI
related also...

But, seeing how this is an ACPI related issue, and ACPI is a "good
thing" for my laptop (keeps it from over heating, lets me know how much
battery I have left, etc.), is there a way to keep the kernel from
loading the synaptics driver so that the touchpad can be treated like a
regular PS/2 mouse?

> The best solution all around, is to have an XFree86 driver
> development kit, which allows externally supplied drivers that do
> not come with XFree86 to be compiled with minimal source code and
> required files.  XFree86 has a package called "XFree86-SDK" which
> is intended to allow drivers to be built externally like this,
> which is minimal in size, however doesn't maintain
> this functionality very well and it is usually broken quite
> badly.  It also doesn't work properly on all architectures we
> need to support, and as such, the SDK is currently unuseable for
> us.  Also, the SDK only works for video drivers currently, and 
> not for input drivers.
> At some point in the future, I will be making attempts to fix the 
> XFree86 SDK, make it work on all architectures properly, and add 
> support for input drivers.

Awesome, I'll volunteer right now to test :)

> This work is not planned to occur in the Cambridge project
> timeframe however, and so the synaptics driver is in limbo right
> now.  I am going to be investigating this more deeply in a few 
> weeks time, however I'm not sure if there will be a sane way for 
> me to include this driver in RPM packaging or not.

Understandable since the 2.6 kernel isn't planned for Cambridge. I plan
to avoid the 2.6 kernel until I can use my laptop without an USB mouse
attached to it.

> Thanks for your attention.

You're welcome

Joshua Legbandt <jtlegbandt at>

More information about the test mailing list