[Gta04-owner] [PATCH] ov9655: fix write_regs loop

[Dr. H. Nikolaus Schaller]

Am 07.06.2013 um 23:54 schrieb Andreas Kemnade:

> The no need to write-optimization in ov9655_write_regs caused the
> loop to abort at the first register which needs not to be changed.
> So many registers were not set.
> 
> Signed-off-by: Andreas Kemnade <andreas at kemnade.info>
> ---
> drivers/media/i2c/ov9655.c |   13 ++++++-------
> 1 files changed, 6 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/media/i2c/ov9655.c b/drivers/media/i2c/ov9655.c
> index 3904d33..7b2859f 100644
> --- a/drivers/media/i2c/ov9655.c
> +++ b/drivers/media/i2c/ov9655.c
> @@ -735,19 +735,18 @@ static int ov9655_write_regs(struct i2c_client *client,
> {
> 	int i, ret;
> 	for (i = 0; i < n; i++) {
> -		u8 val = regs->value;
> -		if(regs->clear != 0) { /* modify only some bits */
> -			ret = ov9655_read(client, regs->addr);
> +		u8 val = regs[i].value;
> +		if(regs[i].clear != 0) { /* modify only some bits */
> +			ret = ov9655_read(client, regs[i].addr);
> 			if(ret < 0)
> 				return ret;
> -			val |= (ret & regs->clear);
> +			val |= (ret & regs[i].clear);
> 			if(val == (u8) ret)
> -				return 0;	/* no need to write */
> +				continue;	/* no need to write */

Oops! You are right. This optimization was introduced very
late and is rarely used. But if it is, it fails.

Thanks for spotting this problem and submitting the patch!

> 		}
> -		ret = ov9655_write(client, regs->addr, val);
> +		ret = ov9655_write(client, regs[i].addr, val);
> 		if (ret < 0)
> 			return ret;
> -		regs++;
> 	}
> 	return 0;
> }
> -- 
> 1.7.2.5
> 
> _______________________________________________
> Gta04-owner mailing list
> Gta04-owner at goldelico.com
> http://lists.goldelico.com/mailman/listinfo/gta04-owner