From mboxrd@z Thu Jan  1 00:00:00 1970
To: Fans of the OS Plan 9 from Bell Labs <9fans@9fans.net>
In-reply-to: Your message of "Sat, 02 Jan 2016 10:06:31 PST."
	<9a649e41e74960f62d15a7e14e9cfa07@mule>
References: <594280736.4698467.1451711534944.JavaMail.yahoo.ref@mail.yahoo.com>
	<594280736.4698467.1451711534944.JavaMail.yahoo@mail.yahoo.com>
	<9a649e41e74960f62d15a7e14e9cfa07@mule>
Date: Sat,  2 Jan 2016 11:58:09 -0800
From: Bakul Shah <bakul@bitblocks.com>
Message-Id: <20160102195809.2EA35B82A@mail.bitblocks.com>
Subject: Re: [9fans] Pi updates
Topicbox-Message-UUID: 7d3b3abc-ead9-11e9-9d60-3106f5b1d025

FWIW, when I was looking at doing a gpio drive I was thinking of
something like the following.

Devices:
    #G/gpio/ctl
    #G/gpio/$pin/ctl
    #G/gpio/$pin/data
    #G/gpio/$pin/event
    ...

    bing -a '#G' /dev

A pin must be configured before use (and only if neither of
its data & event file is open). For example:

    echo 'in rising edge high' > /dev/gpio/10/ctl

Syntax
    reset	# reset to default
    in [async][rising|falling][edge] [high|low] [init $value] [buf $value]
    out [pullup|pulldown] [init $value] [strength $value]
    alt [0..5]	# pi specific

You can use the same syntax to configure multiple pins by
writing to /dev/gpio/ctl and appending pin numbers.

    echo 'out pullup 4 12 17' > /dev/gpio/ctl

You can *gang* multiple pins into a single port. For example:

    echo 'create 100 1 2 5 10' > /dev/gpio/ctl

This creates a new "pin" 100 (number must be >= # of physical
pins) and creates a new entry in /dev/gpio.  This succeeds
only if the pins can actually be ganged (this is device
dependent).  You can now configure the port:

    echo 'in rising edge low' > /dev/gpio/100/ctl

To delete a port

    echo 'delete 100' > /dev/gpio/ctl

All constituent pins revert to their default state.

Reading /dev/gpio/N/data returns a sample of the present value.

[The following needs more work]
Reading /dev/pio/N/event blocks the caller until something
changes.  Returns value and timestamp (in case of "async"
inputs, events are handled in the interrupt handler and
buffered).

Random notes:
An "in" pin has an init value since some pins can be
bidirectional.  "alt" may create other devices as a
side-effect such as SPI, I2C, I2S, PWM, UART etc.

Configs are sticky and persist across device opens, which is
why there is a "reset" command.

While this may be easy to use, its implementation would be
somewhat complex.... Ideally one would break this into a small
core kernel mode driver and a fancier user mode one for
configuring.  I stopped at this point for various reasons and
never got back to it.

On Sat, 02 Jan 2016 10:06:31 PST erik quanstrom <quanstro@quanstro.net> wrote:
> On Fri Jan  1 21:15:03 PST 2016, blstuart@bellsouth.net wrote:
> > On Fri, 1/1/16, erik quanstrom <quanstro@quanstro.net> wrote:
> > > i'm looking @ the gpio interface, and i wonder what the recommended
> > > technique for sampling a pin might be from a shell script?
> >
> > I haven't really used it in shell scripts, but if I were going to do
> > so, I'd probably write up a little utility to take a hex string and
> > a bit number and do the 'and' on it.  Then feed that with the
> > result of dd to get exactly one sample from devgpio.  On the
> > other hand, if one of the usual suspects (e.g. hoc, bc, dc, awk)
> > have some bit-wise operators I'm forgetting, use that.
>
> the (atom) aux/number program does do that; none of the others do
> for fairly fundamental reasons: as awk and hoc are really floating
> point, and bc and dc are mp, and the mp library has no bit operations.
> i believe acid can as well, but only with great pain.
>
> perhaps just using base 2 for the encoding would make life a lot easer. so
>
> 00000000000000000000000000010000000000000000000000000000000000000
>
> for bit 27 active.  with this format there's no limit to the number of bits
> one could represent.
>
> also, mightn't there be some value in presenting the full state of the device
> to allow it to be restored directly from the status file, so perhaps 3 lines
> could be added, one each for up/down/float settings?
>
> 00000000000000000000000000010000000000000000000000000000000000000
> 00000000000000000000000000010000000000000000000000000000000000000
> 00000000000000000000000000000000000000000000000000000000000000000
> 00000000000000000000000000000000000000000000000000000000000000000
>
> might represent bit 27 set, with the pull up resistor active.  (perhaps
> float =E2=89=A1 (pullup | pulldown) =3D=3D 0? and could be elimitated.
>
> i'm sure a little tinkering with this idea can make it a lot more useful.
>
> - erik