Re: [PATCH 1/2] spi: Add driver_override SPI device attribute

[Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>]

Hi Trent,

Thanks for your patch!

On Wed, Sep 19, 2018 at 12:50 AM Trent Piepho <tpiepho@xxxxxxxxxx> wrote:
> This attribute works the same was as the identically named attribute
> for PCI, AMBA, and platform devices.  For reference, see:

Note that it doesn't work exactly the same, due to the device_attach()
(see below).

> commit 3cf385713460 ("ARM: 8256/1: driver coamba: add device binding
> path 'driver_override'")
> commit 3d713e0e382e ("driver core: platform: add device binding path
> 'driver_override'")
> commit 782a985d7af2 ("PCI: Introduce new device binding path using
> pci_dev.driver_override")
>
> If the name of a driver is written to this attribute, then the device
> will bind to the named driver and only the named driver.
>
> The device will bind to the driver even if the driver does not list the
> device in its id table.  This behavior is different than the driver's
> bind attribute, which only allows binding to devices that are listed as
> supported by the driver.
>
> It can be used to bind a generic driver, like spidev, to a device.
>
> Signed-off-by: Trent Piepho <tpiepho@xxxxxxxxxx>

> --- a/drivers/spi/spi.c
> +++ b/drivers/spi/spi.c
> @@ -77,6 +77,54 @@ modalias_show(struct device *dev, struct device_attribute *a, char *buf)
>  }
>  static DEVICE_ATTR_RO(modalias);
>
> +static ssize_t driver_override_store(struct device *dev,
> +                                    struct device_attribute *a,
> +                                    const char *buf, size_t count)
> +{
> +       struct spi_device *spi = to_spi_device(dev);
> +       const char *end = memchr(buf, '\n', count);
> +       const size_t len = end ? end - buf : count;
> +       const char *driver_override, *old;
> +       int ret;
> +
> +       /* We need to keep extra room for a newline when displaying value */
> +       if (len >= (PAGE_SIZE - 1))
> +               return -EINVAL;
> +
> +       driver_override = kstrndup(buf, len, GFP_KERNEL);
> +       if (!driver_override)
> +               return -ENOMEM;
> +
> +       device_lock(dev);
> +       old = spi->driver_override;
> +       if (len) {
> +               spi->driver_override = driver_override;
> +       } else {
> +               /* Emptry string, disable driver override */
> +               spi->driver_override = NULL;
> +               kfree(driver_override);
> +       }
> +       device_unlock(dev);
> +       kfree(old);
> +
> +       /* Attach device to new driver if it's not already attached */
> +       ret = device_attach(dev);

This is something the other buses don't do.  Hence they require the user to
explicitly bind the device by writing the device name to the driver's "bind"
file in sysfs. Looks useful to have for the other buses, too.

> +       if (ret < 0 && ret != -EPROBE_DEFER)
> +               dev_warn(dev, "device attach to '%s' failed (%d)\n",
> +                        spi->driver_override, ret);
> +
> +       return count;
> +}
> +
> +static ssize_t driver_override_show(struct device *dev,
> +                                   struct device_attribute *a, char *buf)
> +{
> +       const struct spi_device *spi = to_spi_device(dev);
> +
> +       return snprintf(buf, PAGE_SIZE, "%s\n", spi->driver_override ? : "");

This needs to be protected by device_lock()/device_unlock(), as the
string may be freed concurrently.

> +}
> +static DEVICE_ATTR_RW(driver_override);
> +
>  #define SPI_STATISTICS_ATTRS(field, file)                              \
>  static ssize_t spi_controller_##field##_show(struct device *dev,       \
>                                              struct device_attribute *attr, \
> @@ -158,6 +206,7 @@ SPI_STATISTICS_SHOW(transfers_split_maxsize, "%lu");
>
>  static struct attribute *spi_dev_attrs[] = {
>         &dev_attr_modalias.attr,
> +       &dev_attr_driver_override.attr,
>         NULL,
>  };
>
> @@ -305,6 +354,10 @@ static int spi_match_device(struct device *dev, struct device_driver *drv)
>         const struct spi_device *spi = to_spi_device(dev);
>         const struct spi_driver *sdrv = to_spi_driver(drv);
>
> +       /* Check override first, and if set, only use the named driver */
> +       if (spi->driver_override)
> +               return strcmp(spi->driver_override, drv->name) == 0;
> +
>         /* Attempt an OF style match */
>         if (of_driver_match_device(dev, drv))
>                 return 1;
> diff --git a/include/linux/spi/spi.h b/include/linux/spi/spi.h
> index 7bb36145e2ba..d90e55029704 100644
> --- a/include/linux/spi/spi.h
> +++ b/include/linux/spi/spi.h
> @@ -168,6 +168,7 @@ struct spi_device {
>         void                    *controller_state;
>         void                    *controller_data;
>         char                    modalias[SPI_NAME_SIZE];
> +       const char              *driver_override;
>         int                     cs_gpio;        /* chip select gpio */
>
>         /* the statistics */

spidev_release() should call kfree(spi->driver_override), to avoid a leak.

The same bug is present in the amba and pci variants. Care to send a fix?

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