Re: [PATCH v2 1/4] gpiolib: Add "unknown" direction support

[Alexander Stein]

On Monday 21 February 2011, 12:22:40 Wolfram Sang wrote:
> On Mon, Feb 21, 2011 at 12:07:27PM +0100, Alexander Stein wrote:
> > On Monday 21 February 2011, 10:47:56 Wolfram Sang wrote:
> > > > We had exported our 5V_enable gpio to sysfs to allow a user-space
> > > > application to enable/disable devices connected to 5V circuit.
> > > > But on the other hand we had to read the current status of this gpio
> > > > in the power-fail interrupt handler to distinguish between
> > > > false-positive (5V disabled) and a correct detection.
> > > 
> > > What about gpio_export() (description in Documentation/gpio.txt)?
> > 
> > Ah, I didn't know about this. I just expected this is only used from
> > sysfs part. But you have to make sure your .ko is loaded before
> > userspace is accessing sysfs and tries to export the GPIO.
> 
> Eh? Userspace doesn't export the GPIO in that case.

Sure, but you have to make sure you have it exported or userspace will fail.

> > Or is it "allowed" by the API convention to gpio_request and gpio_export
> > (and set direction) a GPIO in the machine startup code which will later
> > be used in a different place?
> 
> different place = userspace? Well, that's the main intention of
> gpio_export(). (I have the feeling we are missing each other here,
> thoguh) I'd suggest looking a bit further in the docs/code. It should
> make clear what is possible.

No, with different place I mean a kernel module driver which will be loaded 
later using insmod. In this case this module is just expecting the GPIO has 
already be exported and set proper direction without requesting the GPIO 
itself.

Alexander
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/