Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <1548440422-27943-1-git-send-email-justinpopo6@gmail.com> <20190126183011.00f03cc3@archlinux>
In-Reply-To: <20190126183011.00f03cc3@archlinux>
From:   Justin Chen <justinpopo6@gmail.com>
Date:   Wed, 30 Jan 2019 21:15:45 -0800
Message-ID: <CAJx26kXY+Hg-M11LJ-ATP8y_-8RJ=gQ2DPn2Y9YDJHDkQNh=Kw@mail.gmail.com>
Subject: Re: [PATCH v2] iio: adc: ti-ads7950: inconsistency with spi msg
To:     Jonathan Cameron <jic23@kernel.org>
Cc:     linux-iio@vger.kernel.org, knaack.h@gmx.de, lars@metafoo.de,
        linux-kernel@vger.kernel.org,
        Florian Fainelli <f.fainelli@gmail.com>,
        Mark Brown <broonie@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Sat, Jan 26, 2019 at 10:30 AM Jonathan Cameron <jic23@kernel.org> wrote:
>
> On Fri, 25 Jan 2019 10:20:22 -0800
> justinpopo6@gmail.com wrote:
>
> > From: Justin Chen <justinpopo6@gmail.com>
> >
> > To read a channel we require 3 cycles to send, process, and receive
> > the data. The transfer buffer for the third transaction is left blank.
> > This leaves it up to the SPI driver to decide what to do.
> Interesting.  I think that means you may have a bug in your SPI driver.
> The pointer in question is always left set to NULL in the adc
> driver (not explicitly but it will be because ultimately comes from
> a kzalloc).
>
> Documentation for an SPI message makes it clear that NULL is
> allowed.  From include/linux/spi/spi.h
>
> "* Those segments always read the same number of bits as they
>  * write; but one or the other is easily ignored by passing a null buffer
>  * pointer. "
>
> Hmm. It does say 'ignored'.  My assumption has always been that
> means it would be filled with zeros, but maybe I'm wrong.
>
> Mark?
>
I looked more into this and other SPI drivers do fill it with zeros.
I'll submit a patch for the SPI driver instead and see what the SPI
maintainers say.

Thanks,
Justin
> >
> > In one particular case, if the tx buffer is not set the spi driver
> > sets it to 0xff. This puts the ADC in a alarm programming state,
> > therefore the following read to a channel becomes erroneous.
> >
> > Instead of leaving us to the mercy of the SPI driver, we send the
> > ADC cmd on the third transaction to prevent inconsistent behavior.
> >
> > Fixes: 902c4b2446d4 ("iio: adc: New driver for TI ADS7950 chips")
> > Signed-off-by: Justin Chen <justinpopo6@gmail.com>
> > ---
> >  drivers/iio/adc/ti-ads7950.c | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/iio/adc/ti-ads7950.c b/drivers/iio/adc/ti-ads7950.c
> > index 0ad6359..1255d8b 100644
> > --- a/drivers/iio/adc/ti-ads7950.c
> > +++ b/drivers/iio/adc/ti-ads7950.c
> > @@ -422,6 +422,7 @@ static int ti_ads7950_probe(struct spi_device *spi)
> >       st->scan_single_xfer[1].tx_buf = &st->single_tx;
> >       st->scan_single_xfer[1].len = 2;
> >       st->scan_single_xfer[1].cs_change = 1;
> > +     st->scan_single_xfer[2].tx_buf = &st->single_tx;
> >       st->scan_single_xfer[2].rx_buf = &st->single_rx;
> >       st->scan_single_xfer[2].len = 2;
> >
>