On Tue, Nov 22, 2016 at 09:52:17PM +0800, Gao Pan wrote:
> This patch adds lpspi driver to support new i.MX products which use
> lpspi instead of ecspi.

I'll apply this but there's some small issues, can you please fix them
up with incremental patches:

> +static int fsl_lpspi_prepare_message(struct spi_master *master,
> +				     struct spi_message *msg)
> +{
> +	struct fsl_lpspi_data *fsl_lpspi = spi_master_get_devdata(master);
> +
> +	return clk_prepare_enable(fsl_lpspi->clk);
> +}

Rather than doing this per message it's a bit better to do it in
prepare_transfer_hardware(), that way if there's a sequence of messages
queued one after another we don't turn the clock on and off all the
time.  In practice it's likely to make a negligable difference to
performance but it's just better practice.

> +static int fsl_lpspi_txfifo_empty(struct fsl_lpspi_data *fsl_lpspi)
> +{
> +	u32 txcnt;
> +	unsigned long orig_jiffies = jiffies;
> +
> +	do {
> +		txcnt = readl(fsl_lpspi->base + IMX7ULP_FSR) & 0xff;
> +
> +		if (time_after(jiffies, orig_jiffies + msecs_to_jiffies(500))) {
> +			dev_dbg(fsl_lpspi->dev, "txfifo empty timeout\n");
> +			return -ETIMEDOUT;
> +		}
> +		schedule();

Call cond_resched() instead, it has the effect you're looking for in a
more modern way.