[Libdlo] [ANNOUNCE] displaylink-mod-0.3(alpha) (was udlfb)

Greg KH greg at kroah.com
Thu Jul 9 16:14:21 PDT 2009


On Thu, Jul 09, 2009 at 11:54:53PM +0100, Dan wrote:
> Roberto De Ioris wrote:
> > I do not know why hp and samsung has made this decision, but on linux is
> > a pain in the ass setting a configuration for usb devices as the host
> > controller driver get the most useful one using an (IMHO totally wrong)
> > algorithm.
> >
> > The only (raw) solution i found is to unload the usbhid module, then set
> > the first configuration for the usb device using /sys (look for the file
> > bConfigurationValue, it must contain '1' and not '2') and then reload
> > the displaylink module.
> >
> > I must admit that i have no experience on multiple configuration usb
> > device, so probably someone could point me to a real solution.
> >   
> Bernie's given a very thorough rationale for the basis of this design - 
> it does make a bit harder for linux unfortunately.
> 
> The libdlo code calls usb_detach_kernel_driver_np, this in turn calls 
> http://lxr.linux.no/linux+v2.6.30/drivers/usb/core/devio.c#L1465 .
> 
> Maybe you can use this code? As a VID/PID match your probe call should 
> trump the HID class driver, and when the driver loads it should be able 
> to dislodge an existing driver.

Nope, Linux doesn't work that way :(

We need to patch the hid driver and add these devices to the blacklist.
Hopefully displaylink will provide me a list of the ids to blacklist and
that will solve the problem.

thanks,

greg k-h


More information about the Libdlo mailing list