Re: [PATCH 1/11] MFD: DA9052 MFD core module v1

[Mark Brown <broonie@opensource.wolfsonmicro.com>]

On Tue, Jun 28, 2011 at 07:46:49PM +0530, ashishj3 wrote:

> +static int da9052_add_subdevs(struct da9052 *da9052)
> +{
> +	struct da9052_pdata *pdata = da9052->dev->platform_data;
> +	int ret;
> +
> +	static struct mfd_cell __initdata da9052_subdev_info[] = {
> +		{"da9052-onkey", .resources = &da9052_onkey_resource,
> +		 .num_resources = 1},

It seems a bit odd that this is embedded into the function?

> +		{"da9052-gpio", .resources = NULL, .num_resources = 0},

No need to explicitly initialize static data to zero.

> +int da9052_device_init(struct da9052 *da9052)
> +{
> +	struct da9052_pdata *pdata = da9052->dev->platform_data;
> +	int ret;
> +
> +	mutex_init(&da9052->io_lock);
> +	mutex_init(&da9052->auxadc_lock);
> +	pdata->init(da9052);

This will crash if no init() function is provided which seems wrong,
especially when I'd expect people wouldn't have any need to use such a
callback normally.

> +
> +	ret  = da9052_add_subdevs(da9052);
> +	if (ret != 0)
> +		return ret;
> +
> +	ret = da9052_irq_init(da9052, pdata);
> +	if (ret != 0)
> +		return ret;
> +
> +	return 0;
> +}

This doesn't remove things it added when it failed.

> +	for (raddr = reg ; raddr < reg + bytes; raddr++) {
> +		raddr = (raddr << 1);
> +
> +		spi_message_init(&message);
> +		memset(&xfer, 0, sizeof(xfer));
> +
> +		xfer.len = 2;
> +		xfer.tx_buf = da9052->spi_tx_buf;
> +		xfer.rx_buf = da9052->spi_rx_buf;
> +
> +		da9052->spi_tx_buf[0] = raddr;
> +		da9052->spi_tx_buf[1] = *val++;
> +
> +		spi_message_add_tail(&xfer, &message);
> +
> +		spi_sync(da9052->spi_dev, &message);
> +	}

This looks like an open coded spi_write().

> +	for (raddr = reg ; raddr < reg + bytes; raddr++) {
> +		reg = ((raddr << 1) | da9052->rw_pol);
> +
> +		spi_message_init(&message);
> +		memset(&xfer, 0, sizeof(xfer));
> +
> +		xfer.len = 2;
> +		xfer.tx_buf = da9052->spi_tx_buf;
> +		xfer.rx_buf = da9052->spi_rx_buf;
> +
> +		da9052->spi_tx_buf[0] = raddr;
> +		da9052->spi_tx_buf[1] = 0xff;
> +
> +		da9052->spi_rx_buf[0] = 0;
> +		da9052->spi_rx_buf[1] = 0;
> +
> +		spi_message_add_tail(&xfer, &message);
> +
> +		ret = spi_sync(da9052->spi_dev, &message);
> +
> +		if (ret == 0) {
> +			*val = da9052->spi_rx_buf[1];
> +			val++;
> +			return 0;
> +		}

This looks like an open coded spi_write_then_read(), or even better
spi_w8r8().

> +	da9052_spi->spi_tx_buf = kmalloc(2, GFP_KERNEL | GFP_DMA);
> +	if (!da9052_spi->spi_tx_buf) {
> +		ret = -ENOMEM;
> +		goto err_spi_rx_buf;
> +	}

It would be better to just allocate the array as part of the structure,
a separate allocation just uses more memory for both the pointer and the
blocks that are used for the allocation.

> +static struct spi_driver da9052_spi_driver = {
> +	.probe = da9052_spi_probe,
> +	.remove = __devexit_p(da9052_spi_remove),
> +	. driver = {
> +		.name = "da9052_spi",

Why the _spi?

> index 0000000..c005a28
> --- /dev/null
> +++ b/include/linux/mfd/da9052/da9052.h

> +static const int32_t tbat_lookup[255] = {

This shouldn't be in a header file.  If it needs to be shared between
multiple modules define it in one place and add the prototype in the
header file.