Re: [PATCH 2.6.10-rc2-bk1] media: Update drivers/media/video/arv.c

[Randy.Dunlap]

Hirokazu Takata wrote:
> Hi,
>
> Here is a patch to update AR camera device driver.
> Please apply.
>
> 	* drivers/media/video/arv.c:
> 	- Remove warnings; use module_param() instead of MODULE_PARM(),
>           because MODULE_PARM() is deprecated.
> 	- Fix white-space damages.
>
> Thanks.
>
> Signed-off-by: Hirokazu Takata <takata@xxxxxxxxxxxxxx>
> ---
>
>  arv.c |  118 +++++++++++++++++++++++++++++++++---------------------------------
>  1 files changed, 59 insertions(+), 59 deletions(-)
>
>
> diff -ruNp a/drivers/media/video/arv.c b/drivers/media/video/arv.c
> --- a/drivers/media/video/arv.c	2004-11-15 12:17:04.000000000 +0900
> +++ b/drivers/media/video/arv.c	2004-11-17 15:11:17.000000000 +0900
> @@ -127,12 +127,12 @@ static unsigned char	yuv[MAX_AR_FRAME_BY
>  /* module parameters */
>  /* default frequency */
>  #define DEFAULT_FREQ	50	// 50 or 75 (MHz) is available as BCLK
> -static int freq = DEFAULT_FREQ;	/* BCLK: available 50 or 70 (MHz) */
> +static int freq = DEFAULT_FREQ;	/* BCLK: available 50 or 75 (MHz) */
>  static int vga = 0;		/* default mode(0:QVGA mode, other:VGA mode) */
>  static int vga_interlace = 0;	/* 0 is normal mode for, else interlace mode */
> -MODULE_PARM(freq, "i");
> -MODULE_PARM(vga, "i");
> -MODULE_PARM(vga_interlace, "i");
> +module_param(freq, int, 0644);
> +module_param(vga, bool, 0644);
> +module_param(vga_interlace, bool, 0644);

Do you want freq, vga, and vga_interface to be writeable _after_ the
driver is loaded?  (i.e., dynamic)

> -  	for(i=0; i<1000; i++);
> -  	while( ar_inl(PLDI2CSTS) & PLDI2CSTS_NOACK );
> +	for(i=0; i<1000; i++);
> +	while( ar_inl(PLDI2CSTS) & PLDI2CSTS_NOACK );

That "while( " is still whitespace-damaged.
It should be "while (", with no space before the final ')' also.
similar for several others below here, with "while" or "for".

> -  	/* Trasfer data 1 */
> -  	ar_outl(data1, PLDI2CDATA);
> +	/* Trasfer data 1 */
           Transfer
> +	ar_outl(data1, PLDI2CDATA);

> +	/* Ack wait */
> +	for(i=0; i<1000; i++);
Looks like you need a busy wait or a timed wait here since
the compiler should be (or at least could be) optimizing away
that for loop.
(same for other similar loops below)
Someone else said that a cpu_relax() or barrier() in the for-loop
would also work.

> +	while( ar_inl(PLDI2CSTS) & PLDI2CSTS_NOACK );

I would combine the delay loop (above) with the while condition loop.

> -  	/* Trasfer data 2 */
> -  	ar_outl(data2, PLDI2CDATA);
> +	/* Trasfer data 2 */
           Transfer
> +	ar_outl(data2, PLDI2CDATA);

> +	if(n==3){
        if (n == 3) {
> +		/* Trasfer data 3 */
                   Transfer
> +		ar_outl(data3, PLDI2CDATA);
>  		while ( ar_inl(ARVCR0) & ARVCR0_VDS );    // wait for VSYNC
>  		while ( !(ar_inl(ARVCR0) & ARVCR0_VDS) ); // wait for VSYNC
> -	  	ar_outl(PLDI2CSTEN_STEN, PLDI2CSTEN);
> +		ar_outl(PLDI2CSTEN_STEN, PLDI2CSTEN);

> @@ -233,8 +233,8 @@ static inline void wait_for_vertical_syn
>  	int l;
>  
>  	/*
> - 	 * check HCOUNT because we can not check vertual sync.
> - 	 */
> +	 * check HCOUNT because we can not check vertual sync.
                                   cannot        virtual (or
                                                 vertical ?)
> +	 */

--
~Randy
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/