Sorry, should have read Mark's feedback first :)
James
On 12/08/13 17:10, James Hogan wrote:
> On 09/08/13 19:51, John Crispin wrote:
>> +#ifdef DEBUG
>> +#define spi_debug(args...) printk(args)
>> +#else
>> +#define spi_debug(args...)
>> +#endif
>
> This looks a bit like pr_debug. If you have a device pointer around,
> there's also a dev_dbg which takes an additional device pointer and
> prepends it's name to the message.
>
>> +static int rt2880_spi_probe(struct platform_device *pdev)
>> +{
>
> <snip>
>
>> + clk = devm_clk_get(&pdev->dev, NULL);
>> + if (IS_ERR(clk)) {
>> + dev_err(&pdev->dev, "unable to get SYS clock, err=%d\n",
>> + status);
>> + return PTR_ERR(clk);
>> + }
>
> <snip>
>
>> +}
>> +
>> +static int rt2880_spi_remove(struct platform_device *pdev)
>> +{
>> + struct spi_master *master;
>> + struct rt2880_spi *rs;
>> +
>> + master = dev_get_drvdata(&pdev->dev);
>> + rs = spi_master_get_devdata(master);
>> +
>> + clk_disable(rs->clk);
>> + clk_put(rs->clk);
>
> The devm_clk_get in your probe function means you don't need clk_put here.
>
> Cheers
> James
>
