Re: [RFC 10/12] [media] tc358743: detect chip by ChipID instead of IntMask

["Mats Randgaard (matrandg)" <matrandg@cisco.com>]

On 03/30/2015 01:10 PM, Philipp Zabel wrote:
> When resetting the CPU instead of power cycling, IntMask is still set
> from last boot. Instead of depending on it to be set to its reset
> defaults, check the ChipID register and rewrite IntMask if needed.
>
> Signed-off-by: Philipp Zabel<p.zabel@pengutronix.de>
> ---
>   drivers/media/i2c/tc358743.c | 12 +++++++++---
>   1 file changed, 9 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/media/i2c/tc358743.c b/drivers/media/i2c/tc358743.c
> index 2cf97d9..02b131b 100644
> --- a/drivers/media/i2c/tc358743.c
> +++ b/drivers/media/i2c/tc358743.c
> @@ -1847,13 +1847,19 @@ static int tc358743_probe(struct i2c_client *client,
>   	sd->flags |= V4L2_SUBDEV_FL_HAS_DEVNODE | V4L2_SUBDEV_FL_HAS_EVENTS;
>   
>   	/* i2c access */
> +	if (i2c_rd16(sd, CHIPID) != 0x0000) {
> +		v4l2_info(sd, "not a TC358743 on address 0x%x\n",
> +			  client->addr << 1);
> +		return -ENODEV;
> +	}

Yes, it is better to read a read-only-register.

>   	/* read the interrupt mask register, it should carry the
>   	 * default values, as it hasn't been touched at this point.
>   	 */
>   	if (i2c_rd16(sd, INTMASK) != 0x0400) {
> -		v4l2_info(sd, "not a TC358743 on address 0x%x\n",
> -			  client->addr << 1);
> -		return -ENODEV;
> +		v4l2_warn(sd, "initial interrupt mask: 0x%04x\n",
> +			  i2c_rd16(sd, INTMASK));
> +		i2c_wr16(sd, INTMASK, 0x0400);
> +	}
>   
>   	tc358743_clear_interrupt_status(sd);

Are there other registers that are not reset? If so, I think the driver 
should reset them all during initialization.

Regards,
Mats Randgaard