input-evdev patch adding option to override XDevice type Atom, option to force EV_REL events being treated as absolute values.

Greg KH greg at
Tue Mar 11 09:08:50 PDT 2008

On Tue, Mar 11, 2008 at 04:45:53PM +0200, Daniel Stone wrote:
> On Tue, Mar 11, 2008 at 03:25:26PM +0100, Wolfgang Draxinger wrote:
> > Am Dienstag, 11. M?rz 2008 schrieb Nicolas Mailhot:
> > > I'd be very careful not to assume what can or can not be done at
> > > kernel level, and what should and should not belong in the kernel.
> > 
> > I think applying the quirks at the kernel level in the [usb]hid driver 
> > is probably the way to go, but IMHO having the _table_ mapping the 
> > quirks to devices _stored in_ the kernel is a bad idea. That's what 
> > we've got SysFS and UDev for. I mean, modules are configured on the 
> > user space level as well (/etc/modules.conf, /etc/modprobe.conf), 
> > though they live in kernel space.
> > 
> > Have the table somewhere in the filesystem (possibly in form of UDev 
> > rules) and let UDev apply the quirks on individual devices through 
> > SysFS.
> The point is that you're going against every possible precedent (PS/2
> devices, ATAPI, random USB devices, whatever -- all of which allegedly
> speak 'standard' protocols, but which need dumb vendor quirks).  Tables
> in the kernel are stock standard by now.

I agree, what's one more :)

Is anyone thinking of asking the current HID maintainer about this?  He
would be the best one to have in the conversation...

More information about the xorg mailing list