Re: [PATCH/RFC v2 3/7] spi: core: Add support for registering SPI slave controllers

[Geert Uytterhoeven]

On Mon, Sep 12, 2016 at 10:50 PM, Geert Uytterhoeven
<geert+renesas@xxxxxxxxx> wrote:
> --- a/drivers/spi/spi.c
> +++ b/drivers/spi/spi.c
> @@ -1477,15 +1477,6 @@ static int of_spi_parse_dt(struct spi_master *master, struct spi_device *spi,

> @@ -1799,7 +1908,6 @@ struct spi_master *spi_alloc_master(struct device *dev, unsigned size)
>         device_initialize(&master->dev);
>         master->bus_num = -1;
>         master->num_chipselect = 1;
> -       master->dev.class = &spi_master_class;
>         master->dev.parent = dev;
>         pm_suspend_ignore_children(&master->dev, true);
>         spi_master_set_devdata(master, &master[1]);
> @@ -1882,9 +1990,18 @@ int spi_register_master(struct spi_master *master)
>         if (!dev)
>                 return -ENODEV;
>
> -       status = of_spi_register_master(master);
> -       if (status)
> -               return status;
> +       if (master->flags & SPI_CONTROLLER_IS_SLAVE) {
> +               if (!IS_ENABLED(CONFIG_SPI_SLAVE))
> +                       return -ENOSYS;
> +
> +               master->dev.class = &spi_slave_class;
> +       } else {
> +               master->dev.class = &spi_master_class;
> +
> +               status = of_spi_register_master(master);
> +               if (status)
> +                       return status;
> +       }
>
>         /* even if it's just one always-selected device, there must
>          * be at least one chipselect

0day reported a warning during boot:

    WARNING: CPU: 0 PID: 1 at drivers/base/core.c:251 device_release+0x7d/0x8a
    Device '(null)' does not have a release() function, it is broken
and must be fixed.

This is caused by moving the setup of master->dev.class.
To fix this, I can
  1) Introduce a separate spi_alloc_slave() function, which sets up
     spi_slave_class instead of spi_master class,
  OR
  2) Keep the setup of spi_master_class in spi_alloc_master(), and override it
     later. Both classes use the same dev_release method anyway.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds