[PATCH] spi/pl022: make sure the chip_info dev pointer is set

Grant Likely grant.likely at secretlab.ca
Thu Sep 30 10:33:43 EDT 2010


On Thu, Sep 30, 2010 at 6:40 PM, Linus Walleij
<linus.walleij at stericsson.com> wrote:
> We forgot to set the pointer back to the device if the user
> supplied all device settings, leading to crashing debug prints
> as controller data is verified.
>
> Signed-off-by: Linus Walleij <linus.walleij at stericsson.com>
> ---
>  drivers/spi/amba-pl022.c |    2 ++
>  1 files changed, 2 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/spi/amba-pl022.c b/drivers/spi/amba-pl022.c
> index 4c37c4e..90a7e5d 100644
> --- a/drivers/spi/amba-pl022.c
> +++ b/drivers/spi/amba-pl022.c
> @@ -1574,6 +1574,8 @@ static int pl022_setup(struct spi_device *spi)
>                chip_info->duplex = SSP_MICROWIRE_CHANNEL_FULL_DUPLEX;
>                chip_info->cs_control = null_cs_control;
>        } else {
> +               /* Users should not set this field */
> +               chip_info->dev = &spi->dev;
>                dev_dbg(&spi->dev,
>                        "using user supplied controller_data settings\n");
>        }

Nack.  data passed to a driver via the controller_data must be
immutable as far as the device driver is concerned.  I realize that
this has been rather sloppy in the past; but I'm enforcing it now.
controller_data is static data passed to the driver, and the 'dev'
member of pl022_config_chip needs to be removed.

So, instead of setting the dev pointer here; this code should *remove*
the dev pointer from the structure and use spi->dev directly instead.
I've looked at the driver, and I don't see any significant locations
using chip_info that cannot get the chip_info directly from the struct
spi_device.

g.


> --
> 1.6.3.3
>
>



-- 
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.



More information about the linux-arm-kernel mailing list