Re: [PATCH v4 5/6] spi: at91-usart: add driver for at91-usart as spi

[Radu Pirea]

On 05/28/2018 11:21 AM, Andy Shevchenko wrote:
> On Fri, May 25, 2018 at 8:19 PM, Radu Pirea <radu.pirea@xxxxxxxxxxxxx> wrote:
> > This is the driver for at91-usart in spi mode. The USART IP can be configured
> > to work in many modes and one of them is SPI.
> >
> > The driver was tested on sama5d3-xplained and sama5d4-xplained boards with
> > enc28j60 ethernet controller as slave.
>
> > +#include <linux/of_gpio.h>
>
> What is the use of it?

I need of_gpio.h for of_gpio_named_count, of_get_named_gpio and 
devm_gpio_request_one(found in gpio.h)

> > +#define US_INIT                        (US_MR_SPI_MASTER | US_MR_CHRL | US_MR_CLKO | \
> > +                               US_MR_WRDBT)
>
> Don't split lines like this, it's hard to read.
>
> #define FOO \
>   (BAR1 | BAR2)

I'll fix it.

> I think I already told this to someone recently, maybe to you.
>
> > +/* Register access macros */
> > +#define spi_readl(port, reg) \
> > +       readl_relaxed((port)->regs + US_##reg)
> > +#define spi_writel(port, reg, value) \
> > +       writel_relaxed((value), (port)->regs + US_##reg)
> > +
> > +#define spi_readb(port, reg) \
> > +       readb_relaxed((port)->regs + US_##reg)
> > +#define spi_writeb(port, reg, value) \
> > +       writeb_relaxed((value), (port)->regs + US_##reg)
>
> Names are too generic. You better to use the same prefix as for the
> rest, i.e. at91_spi_

Good ideea. I will change the names.

> > +       /*used in interrupt to protect data reading*/
>
> Comment style.
>
> You need to read some existing code, perhaps, to see how it's done.

Ok. I will add the comment.

> > +static inline void at91_usart_spi_tx(struct at91_usart_spi *aus)
> > +{
> > +       unsigned int len = aus->current_transfer->len;
> > +       unsigned int remaining = aus->current_tx_remaining_bytes;
> > +       const u8  *tx_buf = aus->current_transfer->tx_buf;
> > +
>
> > +       if (remaining)
> > +               if (at91_usart_spi_tx_ready(aus)) {
>
> if (x) {
>   if (y) {
> ...
>   }
> }
>
> is equivalent to if (x && y) {}.
>
> Though, considering your intention here, I would rather go with better
> pattern, i.e.
>
> if (!remaining)
>   return;

Thank for suggestion. I will change.

> > +                       spi_writeb(aus, THR, tx_buf[len - remaining]);
> > +                       aus->current_tx_remaining_bytes--;
> > +               }
> > +}
> > +
> > +static inline void at91_usart_spi_rx(struct at91_usart_spi *aus)
> > +{
>
> > +       if (remaining) {
> > +               rx_buf[len - remaining] = spi_readb(aus, RHR);
> > +               aus->current_rx_remaining_bytes--;
> > +       }
>
> Ditto.
>
> > +}
>
>
> > +static int at91_usart_gpio_setup(struct platform_device *pdev)
> > +{
>
> > +       struct device_node      *np = pdev->dev.parent->of_node;
>
> Your driver is not OF specific as far as I can see. Drop all these
> device_node stuff and change API calls respectively.

Ok. What do you suggest to use instead of OF API to get the count of 
cs-gpios and to read their values one by one?

> > +       int                     i;
>
> > +       int                     ret = 0;
> > +       int                     nb = 0;
>
> What happened to indentation?
>
> Redundnant assignment for both.
>
> > +       if (!np)
> > +               return -EINVAL;
> > +
> > +       nb = of_gpio_named_count(np, "cs-gpios");
> > +       for (i = 0; i < nb; i++) {
> > +               int cs_gpio = of_get_named_gpio(np, "cs-gpios", i);
> > +
> > +               if (cs_gpio < 0)
> > +                       return cs_gpio;
> > +
> > +               if (gpio_is_valid(cs_gpio)) {
> > +                       ret = devm_gpio_request_one(&pdev->dev, cs_gpio,
> > +                                                   GPIOF_DIR_OUT,
> > +                                                   dev_name(&pdev->dev));
> > +                       if (ret)
> > +                               return ret;
> > +               }
> > +       }
> > +
> > +       return 0;
> > +}
> > +
> > +static int at91_usart_spi_probe(struct platform_device *pdev)
> > +{
>
> > +       regs = platform_get_resource(to_platform_device(pdev->dev.parent),
> > +                                    IORESOURCE_MEM, 0);
> > +       if (!regs)
> > +               return -EINVAL;
>
> This looks weird. Supply resource to _this_ device in your MFD code.

I know weird, but is the safest way to pass the resource and the of_node.

> > +       dev_info(&pdev->dev,
> > +                "Atmel USART SPI Controller version 0x%x at 0x%08lx (irq %d)\n",
> > +                spi_readl(aus, VERSION),
> > +                (unsigned long)regs->start, irq);
>
> I think I already told you, don't use explicit casting when print.
> If it wasn't you, do you homework then. But above is no go. >
> > +       return 0;
>
> > +static struct platform_driver at91_usart_spi_driver = {
> > +       .driver = {
> > +               .name = "at91_usart_spi",
>
> > +               .of_match_table = of_match_ptr(at91_usart_spi_dt_ids),
>
> Drop of_match_ptr(). It's not needed.
>
> > +       },
> > +       .probe = at91_usart_spi_probe,
>
> > +       .remove = at91_usart_spi_remove, };
>
> Already told ya, split lines correctly.