On Tue, Jul 27, 2010 at 01:16:33AM -0700, Dmitry Torokhov wrote:

On Mon, Jul 26, 2010 at 03:49:42PM -0700, Greg KH wrote:

quoted

On Fri, Jul 23, 2010 at 09:43:17AM -0700, Dmitry Torokhov wrote:

quoted

On Fri, Jul 23, 2010 at 04:38:59PM +0100, Alan Cox wrote:

quoted

quoted

I am confused here... Are you having a separate SPI driver create a
platform device and then you have mrstouch to bind to this
intermediate platform device? Are you doing that so you can introduce
I2C interface later? If so I think I prefer how adxl34x and ad7879
drivers are structured - they are split into core and interface parts
but do not require extra devices/drivers (see in my 'next' brnach).

There is no SPI interface to the device. It ended up in the kernel SPI
because old versions of the device firmware listed it in the firmware
tables as SPI and rather than doing the right thing (correcting the
type) the x86 code created an SPI device for it.

At a certain level it may be SPI, but it's all hidden behind the
firmware on the SCU and nothing to do with Linux SPI at all.

Ah, OK then.

So, what is the plan of action? Greg, are you going to apply the patches
to staging and move over to input or shoudl I do that (apply Alan's
patches to staging _in my tree_ and move over to drivers.input)?

I already have the driver in my staging-next tree, so I can apply these
6 patches as well, and then move it to the drivers/input subdirectory if
you want.

Or I can leave it all in drivers/staging/ and then after 2.6.36-rc1
comes, you can move the driver into drivers/input/ subdir.

The 2nd one I think is the best.

Great.  I'll queue up these patches then.  Alan, when 2.6.36-rc1 is out,
ask Dmitry to move the driver into the input subdir and all should be
good.

thanks,

greg k-h