Re: [PATCH] Staging: comedi: fix coding style issues in unioxx5.c

[Ian Abbott <abbotti@mev.co.uk>]

On 2013-11-17 03:42, Conrad Gomes wrote:
> Hi,
>
> No problem, I'll break this up and resend it.
>
> In order to use dev_info type functions I was planning on assigning the
> comedi device's struct device to the struct subdevice's class_dev as
> follows in comedi_alloc_subdevices as it appears that is is not used(I
> could be wrong):
>
> diff --git a/drivers/staging/comedi/drivers.c
> b/drivers/staging/comedi/drivers.c
> index 8f02bf6..e4c44ae 100644
> --- a/drivers/staging/comedi/drivers.c
> +++ b/drivers/staging/comedi/drivers.c
> @@ -90,6 +90,7 @@ int comedi_alloc_subdevices(struct comedi_device *dev,
> int num_subdevices)
>                  s->async_dma_dir = DMA_NONE;
>                  spin_lock_init(&s->spin_lock);
>                  s->minor = -1;
> +               s->class_dev = dev->class_dev;
>          }
>          return 0;
>   }

The class_dev member is assigned by comedi_alloc_subdevice_minor(s) 
(which will be called if the subdevice supports asynchronous comedi 
commands) and is kept around for the call to comedi_free_subdevice_minor(s).

The struct comedi_subdevice has a pointer back to the struct 
comedi_device.  The class_dev member in the struct comedi_device can be 
used as the device pointer parameter of dev_info() etc.

> And adding a reference to the subdevice in the private  data structure
> unioxx5_subd_priv in order to substitute pr_info with dev_info. If this
> is fine I'll remove all references of pr_info in the file:
>
> diff --git a/drivers/staging/comedi/drivers/unioxx5.c
> b/drivers/staging/comedi/drivers/unioxx5.c
> index 2b1ece8..bb8f5a4 100644
> --- a/drivers/staging/comedi/drivers/unioxx5.c
> +++ b/drivers/staging/comedi/drivers/unioxx5.c
> @@ -72,6 +72,7 @@ Devices: [Fastwel] UNIOxx-5 (unioxx5),
>
>   /* 'private' structure for each subdevice */
>   struct unioxx5_subd_priv {
> +       struct comedi_subdevice *subdevice;
>          int usp_iobase;
>          /* 12 modules. each can be 70L or 73L */
>          unsigned char usp_module_type[12];
> @@ -94,9 +95,10 @@ static int __unioxx5_define_chan_offset(int chan_num)
>   static void __unioxx5_digital_config(struct unioxx5_subd_priv *usp,
> int mode)
>   {
>          int i, mask;
> +       struct device *dev = usp->subdevice->class_dev;
>
>          mask = (mode == ALL_2_OUTPUT) ? 0xFF : 0x00;
> -       printk("COMEDI: mode = %d\n", mask);
> +       dev_info(dev, "COMEDI: mode = %d\n", mask);
>
>          outb(1, usp->usp_iobase + 0);

It would be better to change those functions to pass the pointer to the 
struct comedi_subdevice instead of the pointer to the struct 
unioxx5_subd_priv.

A variable or parameter named 'dev' in the comedi drivers is usually of 
type struct comedi_device for consistency.

The message passed to dev_info() wouldn't need the "COMEDI:" prefix - 
it's redundant.

Also, this particular message should be a dev_dbg() at best, and 
probably better not to have it at all!

For the sake of example, the above function could start as follows:

static void __unioxx5_digital_config(struct comedi_subdevice *s, int mode)
{
	struct unioxx5_subd_priv *usp = s->private;
	struct device *csdev = s->device->class_dev;
	int i, mask;

	mask = (mode == ALL_2_OUTPUT) ? 0xFF : 0x00;
	dev_dbg(csdev, "mode = %d\n", mask);

-- 
-=( Ian Abbott @ MEV Ltd.    E-mail: <abbotti@mev.co.uk>        )=-
-=( Tel: +44 (0)161 477 1898   FAX: +44 (0)161 718 3587         )=-