Re: [rtc-linux] [PATCH] rtc: add support for maxim dallas rtc ds1343 and ds1344

[Mark Brown <broonie@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 1464 bytes --]

On Sun, Apr 13, 2014 at 08:12:57PM +0530, RAGHAVENDRA GANIGA wrote:

> +static const struct spi_device_id ds1343_id[] = {
> +	{ "ds1343", 0 },
> +	{ "ds1344", 1 },
> +	{ }
> +};
> +MODULE_DEVICE_TABLE(spi, ds1343_id);

If the 0 and 1 mean anything there they should have #defines, otherwise
just omit them.

> +static int ds1343_get_reg(struct device *dev, unsigned char address,
> +				unsigned char *buf)
> +{
> +	struct spi_device *spi = to_spi_device(dev);
> +
> +	/* MSB of the spi address                        _
> +	   in this rtc should be zero for read operation R/W */
> +	*buf = address;
> +
> +	return spi_write_then_read(spi, buf, 1, buf, 1);
> +}

This and the set_reg() function look like you should be using regmap.

> +static int ds1343_ioctl(struct device *dev, unsigned int cmd, unsigned long arg)
> +{
> +	switch (cmd) {
> +#ifdef RTC_SET_CHARGE
> +	case RTC_SET_CHARGE:
> +	{
> +		int val;
> +
> +		if (copy_from_user(&val, (int __user *)arg, sizeof(int)))
> +			return -EFAULT;
> +
> +		return ds1343_set_reg(dev, DS1343_TRICKLE_REG, val);
> +	}
> +	break;
> +#endif
> +	}

What defines this?  I notice that ds1302 also does this - is this device
different enouh to need a separate driver?

> +static irqreturn_t ds1343_irq(int irq, void *dev_id)
> +{
> +	struct ds1343_priv *priv = dev_id;
> +
> +	disable_irq_nosync(irq);
> +	schedule_work(&priv->work);
> +	return IRQ_HANDLED;
> +}

Use request_threaded_irq() rather than open coding it.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]