Re: [patch/rfc 2.6.25-git] gpio: sysfs interface

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

On Tuesday 29 April 2008, Trent Piepho wrote:
> > Sorry if I'm being dense; how do you want this bit to work?  As I see
> > it, there are a few options:
> >
> > 1) Have the files named as you suggest and all of them always present,
> > albeit read-only until export.  Very easy to use, easy to discover which
> > file is which, a decent bit of memory usage having them all listed.
> 
> Well, is it really that much?  There are 579 files under /sys/class/tty.  But
> suppose it is too much (why isn't tty too much then?), then we can do 3.

I just ssh'd into three embedded boards I have handy, and they have
respectively four, four, and seven entries there.  That "seven"
case is actually incorrect ... the other three serial ports aren't
connected to anything.

So:  yes, adding a few hundred useless sysfs nodes *IS* a problem
in the target environment of embedded boards.

Note that "read-only until export" is far from straightforward
to achieve.


> > 3) Have the files named as you suggest, explicit export/request but
> > better parsing behind the control file so something like
> > echo "export pca9557-0:5" > control
> > works.  Very very nice for the user, big heavy back end.
> 
> The back end doesn't seem that big to me.  Here's code for it.

Which fails in a common case:  chip labels are not unique.


> If anything, 
> the parsing code is simpler than what David has.

Apples vs oranges.  Use the same command syntax if you're going
to make comparisons; I can save even more with "+export/-unexport"
syntax.  For comparable syntax, your stuff *IS* bigger.


> David's code for parsing the control file plus code for generating a mapping
> range file would certainly be larger.

The #3 option presumes some file listing chips and ranges too,
since GPIOs are exported only on demand.  Ditto #2 and #4...

- Dave