Re: [patch/rfc 1/4] GPIO implementation framework

[David Brownell <david-b@pacbell.net>]

Thanks, I'll be looking at it ... the one thing I strongly dislike is
this change:

On Tuesday 27 November 2007, eric miao wrote:
> 4. use a loop for "gpio_desc[]" instead of a loop for "gpio_chips[]" in
>    gpiolib_show(), change is straight forward; since it is now per gpio
>    based, the gpio_chip.dbg_show() is removed as well

That removes the ability to display all kinds of significant
chip-specific state ... like whether a given signal has active
pullups or pulldowns, uses open-drain signaling, and so forth.

It also makes access a lot slower ... e.g. rather than one
batch of I2C or SPI operations for all N signals on a chip,
it's got to do N batches.

Plus it just needlessly breaks existing code.

I'll put together a version without that problem.

- Dave