linux-kernel - Re: [PATCH v4 1/3] iio: hx711: add clock-frequency property in DT

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAL_JsqLGn=iTbD+qDasyNN8qpPzsTWp0ZQSK68GaDPx1OBG9xw@mail.gmail.com>
Date:   Fri, 6 Jul 2018 11:38:41 -0600
From:   Rob Herring <robh@...nel.org>
To:     Andreas Klinger <ak@...klinger.de>
Cc:     Jonathan Cameron <jic23@...nel.org>,
        Hartmut Knaack <knaack.h@....de>,
        Lars-Peter Clausen <lars@...afoo.de>,
        Peter Meerwald <pmeerw@...erw.net>,
        Mark Rutland <mark.rutland@....com>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        David Miller <davem@...emloft.net>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Linus Walleij <linus.walleij@...aro.org>,
        Randy Dunlap <rdunlap@...radead.org>,
        devicetree@...r.kernel.org, linux-iio@...r.kernel.org,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v4 1/3] iio: hx711: add clock-frequency property in DT

On Fri, Jul 6, 2018 at 12:22 AM Andreas Klinger <ak@...klinger.de> wrote:
>
> Hi Rob,
>
> please see explanation below. Thanks.
>
> Rob Herring <robh@...nel.org> schrieb am Thu, 05. Jul 15:30:
> > On Wed, Jul 04, 2018 at 02:36:38PM +0200, Andreas Klinger wrote:
> > > Add clock-frequency property for hx711 ADC
> > >
> > > This is the frequency of PD_SCK. It affects only the high value duration
> > > since low value duration is not relevant and we are not able to switch
> > > faster than the minimum duration specified.
> > >
> > > After PD_SCK goes high DOUT is read just before PD_SCK goes down again.
> > > This is necessary because of parasitic capacities on the wiring.
> >
> > s/capacities/capacitance/
>
> I'll change.
>
> >
> > >
> > > Signed-off-by: Andreas Klinger <ak@...klinger.de>
> > > ---
> > >  Documentation/devicetree/bindings/iio/adc/avia-hx711.txt | 10 ++++++++++
> > >  1 file changed, 10 insertions(+)
> > >
> > > diff --git a/Documentation/devicetree/bindings/iio/adc/avia-hx711.txt b/Documentation/devicetree/bindings/iio/adc/avia-hx711.txt
> > > index b3629405f568..4bee51d536e1 100644
> > > --- a/Documentation/devicetree/bindings/iio/adc/avia-hx711.txt
> > > +++ b/Documentation/devicetree/bindings/iio/adc/avia-hx711.txt
> > > @@ -8,11 +8,21 @@ Required properties:
> > >             See Documentation/devicetree/bindings/gpio/gpio.txt
> > >   - avdd-supply:    Definition of the regulator used as analog supply
> > >
> > > +Optional properties:
> > > + - clock-frequency:        Frequency of PD_SCK
> > > +                   This setting affects the duration of the high value
> > > +                   phase of the clock (PD_SCK). The low value phase is
> > > +                   not affected since it is not relevant for the
> > > +                   measurement.
> >
> > That's not how frequency works. The high time should be equal to the low
> > time. If you have parasitic capacitance affecting the rise time, then it
> > is going to affect the fall time too.
> >
> > Perhaps there could be some reason not to have a square wave, but I
> > didn't see one in the datasheet.
>
> When PD_SCK goes high the ADC sets DOUT to the corresponding value. Only after a
> rising edge DOUT is set. The falling edge doesn't matter.

When bit-banging yes, but the simplest h/w implementation would shift
data on one edge and sample on the other.

> The parasitic capacitance is occuring only on DOUT line, but not on PD_SCK. The
> difference is that PD_SCK is driven by the microcontroller, and DOUT by the ADC.

Maybe for you, but for someone else's design? I just don't want to see
the binding evolving fixing one problem at a time.

> So after i set the rising edge on PD_SCK i have to wait until DOUT reaches its
> value. Datasheet says this can take at maximum 0.1 us. This might be true, but
> without wiring between ADC and microcontroller. With wiring of a 10 m cable i
> saw 1 us on the oscilloscope.
>
> If required i can post a screenshot of the oscilloscope.
>
> The code which is executed in the driver on PD_SCK during low value takes much
> more than the requested 0.2 us (measured 0.56 us). Of course i can make it a
> square wave with equal low and high time, but i don't see it's needed and it'll
> slow down the communication speed.
>
> Out of my frequency generator i can get a rectangular waveform with different
> low and high times and it's still called frequency. A don't have another
> expression for it.

Yes, but the frequency is not 1/(high time) as you defined. But I
wasn't expecting 2 orders of magnitude difference.

> Do you have a suggestion how to call this attribute in the DT if frequency is
> not suitable?

If you really want to not have 50% duty cycle, then of course you
would need to define an additional property.

> >
> > > +                   Minimum value allowed is 20 kHz because of maximum
> > > +                   high time of 50 microseconds.
> >
> > So 10kHz should be minimum.
> >
>
> Because the low time is very short (0.5 us) in comparison to the high time
> (50 us) in the actual waveform i neglected it.

Okay, that's not obvious.

Rob