Re: [PATCH] davinci: dm644x: fix enum ccdc_gama_width and enum ccdc_data_size comparision warning

[Mauro Carvalho Chehab <mchehab@infradead.org>]

Em Wed,  2 Jan 2013 17:23:50 +0530
"Lad, Prabhakar" <prabhakar.csengg@gmail.com> escreveu:

> while the effect is harmless this patch

I disagree that this is a harmless warning. It is here for a reason:
you should not be relying on the enum "magic" value, since the main
reason to use an enum is to fill/compare the enum fields only by their
names, and not by their number.

> fixes following build warning,
> 
> drivers/media/platform/davinci/dm644x_ccdc.c: In function ‘validate_ccdc_param’:
> drivers/media/platform/davinci/dm644x_ccdc.c:233:32: warning: comparison between
> ‘enum ccdc_gama_width’ and ‘enum ccdc_data_size’ [-Wenum-compare]
> 
> Signed-off-by: Lad, Prabhakar <prabhakar.lad@ti.com>
> ---
>  drivers/media/platform/davinci/dm644x_ccdc.c |    5 ++++-
>  1 files changed, 4 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/media/platform/davinci/dm644x_ccdc.c b/drivers/media/platform/davinci/dm644x_ccdc.c
> index ee7942b..42b473a 100644
> --- a/drivers/media/platform/davinci/dm644x_ccdc.c
> +++ b/drivers/media/platform/davinci/dm644x_ccdc.c
> @@ -228,9 +228,12 @@ static void ccdc_readregs(void)
>  static int validate_ccdc_param(struct ccdc_config_params_raw *ccdcparam)
>  {
>  	if (ccdcparam->alaw.enable) {
> +		u32 gama_wd = ccdcparam->alaw.gama_wd;
> +		u32 data_sz = ccdcparam->data_sz;
> +
>  		if ((ccdcparam->alaw.gama_wd > CCDC_GAMMA_BITS_09_0) ||
>  		    (ccdcparam->alaw.gama_wd < CCDC_GAMMA_BITS_15_6) ||
> -		    (ccdcparam->alaw.gama_wd < ccdcparam->data_sz)) {
> +		    (gama_wd < data_sz)) {

hmm... from include/media/davinci/dm644x_ccdc.h:
enum ccdc_gama_width {
	CCDC_GAMMA_BITS_15_6,	// 0
	CCDC_GAMMA_BITS_14_5,	// 1
	CCDC_GAMMA_BITS_13_4,	// 2
	CCDC_GAMMA_BITS_12_3,	// 3
	CCDC_GAMMA_BITS_11_2,	// 4
	CCDC_GAMMA_BITS_10_1,	// 5
	CCDC_GAMMA_BITS_09_0	// 6
};

enum ccdc_data_size {
	CCDC_DATA_16BITS,	// 0
	CCDC_DATA_15BITS,	// 1
	CCDC_DATA_14BITS,	// 2
	CCDC_DATA_13BITS,	// 3
	CCDC_DATA_12BITS,	// 4
	CCDC_DATA_11BITS,	// 5
	CCDC_DATA_10BITS,	// 6
	CCDC_DATA_8BITS		// 7
};

That doesn't seem right, as comparing the enum integer value won't
warrant that the number of bits of gamma. For example, gamma == 6
means 9 bits, while ccdc == 6 means 10 bits.

In any case, the code is just crappy, as one could anytime add more
values at the enum or reorder.

So, a better fix would be to have an array that would convert from the
enum "magic" number into the number of bits.

Hmm... wait a moment: why are you using an enum here at the first place???

It seems that it would be a way better to just use 2 unsigned integers:
ccdc_data_num_bits and ccdc_gama_num_bits, and just fill it with the
number of bits, instead of declaring an enum for it.

Another alternative would be to merge them into just one enum, like:

enum ccdc_bits {
	CCDC_8_BITS = 8,
	CCDC_9_BITS = 9,
	CCDC_10_BITS = 10,
	CCDC_11_BITS = 11,
	CCDC_12_BITS = 12,
	CCDC_13_BITS = 13,
	CCDC_14_BITS = 14,
	CCDC_15_BITS = 15,
	CCDC_16_BITS = 16,
};

and replace all occurrences of ccdc_data_size and ccdc_gama_width by
the new enum.

This way, you could trust on compare one field with the other.

>  			dev_dbg(ccdc_cfg.dev, "\nInvalid data line select");
>  			return -1;
>  		}

Regards,
Mauro