[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20220820130248.7773a6b3@jic23-huawei>
Date: Sat, 20 Aug 2022 13:02:48 +0100
From: Jonathan Cameron <jic23@...nel.org>
To: Andy Shevchenko <andy.shevchenko@...il.com>
Cc: Eddie James <eajames@...ux.ibm.com>,
Lars-Peter Clausen <lars@...afoo.de>,
linux-iio <linux-iio@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Joel Stanley <joel@....id.au>
Subject: Re: [PATCH v5 2/2] iio: pressure: dps310: Reset chip if MEAS_CFG is
corrupt
On Thu, 18 Aug 2022 23:16:55 +0300
Andy Shevchenko <andy.shevchenko@...il.com> wrote:
> On Mon, Aug 15, 2022 at 5:57 PM Eddie James <eajames@...ux.ibm.com> wrote:
> >
> > Corruption of the MEAS_CFG register has been observed soon after
> > system boot. In order to recover this scenario, check MEAS_CFG if
> > measurement isn't ready, and if it's incorrect, reset the DPS310
> > and execute the startup procedure.
>
> ...
>
> > + * Called with lock held. Returns a negative value on error, a positive value
> > + * when the device is not ready, and zero when the device is ready.
>
> Can we have
>
> #define DPS310_DEVICE_NOT_READY 1
>
> (or anonymous enum) and return it instead of abstract 1 or any other
> positive number?
Perhaps make it even clearer by returning the need to wait via another parameter
rather than this being (probably) the only place in driver with a postive rc.
bool reset_done;
rc = dps310_check_reset_meas_cfg(data, DPS310_TMP_RDY, &reset_done);
if (reset_done) {
}
>
> ...
>
> > + rc = dps310_check_reset_meas_cfg(data, DPS310_TMP_RDY);
> > if (rc < 0)
> > goto done;
> >
> > + if (rc > 0) {
> > + rate = dps310_get_temp_samp_freq(data);
> > + timeout = DPS310_POLL_TIMEOUT_US(rate);
> > +
> > + /*
> > + * Poll for sensor readiness; base the timeout upon the sample
> > + * rate.
> > + */
> > + rc = regmap_read_poll_timeout(data->regmap, DPS310_MEAS_CFG,
> > + ready, ready & DPS310_TMP_RDY,
> > + DPS310_POLL_SLEEP_US(timeout),
> > + timeout);
> > + if (rc)
> > + goto done;
> > + }
>
> But have you tried to make a helper that takes a pointer to the
> respective function?
>
Powered by blists - more mailing lists