Thread (8 messages) 8 messages, 4 authors, 2010-11-10

Re: [PATCH V3 1/2] input: CMA3000 Accelerometer driver

From: Jonathan Cameron <hidden>
Date: 2010-09-21 10:21:57
Also in: linux-omap, lkml

On 09/21/10 06:46, Hemanth V wrote:
----- Original Message ----- From: "Jonathan Cameron" [off-list ref]
To: "Dmitry Torokhov" <dmitry.torokhov@gmail.com>
Cc: "Hemanth V" <redacted>; <redacted>; <redacted>; <redacted>
Sent: Tuesday, September 14, 2010 6:40 PM
Subject: Re: [PATCH V3 1/2] input: CMA3000 Accelerometer driver

quoted
On 09/14/10 09:00, Dmitry Torokhov wrote:
quoted
On Wed, Sep 08, 2010 at 12:37:40PM +0100, Jonathan Cameron wrote:
quoted
I'm happy to see your driver go in as it is currently, what bothers me is that we will end
up with incompatible interfaces for all the accelerometers Dmitry takes!
This is a valid concern. How about I chicken out and will not merge any
new sysfs knobs until you guys decide on reasonable interface? Most set
up is done by board code anyways, sysfs is more of nice-to-have?
Hemanth, would removing the sysfs hooks from this driver be ok with you?
I'd personally have favoured a merge, add new interfaces when agreed and
deprecate old ones approach, but it is Dmitry's call.  Perhaps it is better
to get the majority of the device functionality in place now and add
the bells and whistles later.

For input device things are probably mostly fixed for a particular board
design.  There are definitely interesting things one can do if the knobs
are available but they (I think) mostly fall outside of using the device
for input!
Dmitry, Jonathan

I am ok to remove the sysfs entries for now, but would cause some limitations
like not being able to change sampling frequency / disabling interrupts runtime.

Wanted to clarify if the intention is to come up with a standard sysfs interface
for all accelerometer drivers under input/misc including adxl34x.
The intent is to come up with an interface covering a much wider range of devices
if at all possible. I've proposed options for conventional threshold and rate of
change interrupts on IIO (you were cc'd IIRC). I'll propose those on to lkml shortly
and suggest that others suggest the interfaces they would like to see added (or
object to the syntax for the ones I've covered!)

To be honest, sampling frequency is much more general and I would imagine applies to
lots of input drivers.  Do you already have a standard for this Dmitry?

Jonathan
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help