Re: i2c: Re: #2 (Re: More modaliases + patchtool))

[Mark Brown <broonie@opensource.wolfsonmicro.com>]

On Fri, Sep 17, 2010 at 04:35:33PM +0300, Dzianis Kahanovich wrote:
> ...
> + 1 attach for "next" branch.

Again, please try to follow the process in SubmittingPatches for sending
patches.

> I unsure in requrement of i2c modaliases, may be this bus undetectable, but I
> not play with special tools with vendors exclude VIA.

I2C is not probeable but drivers can be autoloaded since boards must
register devices before they can be probed.

> --- linux-2.6.36-rc4-git3/sound/soc/s3c24xx/neo1973_wm8753.c	2010-08-02 01:11:14.000000000 +0300
> +++ b/sound/soc/s3c24xx/neo1973_wm8753.c	2010-09-17 12:39:07.917253497 +0300
> @@ -698,6 +698,9 @@ static void __exit neo1973_exit(void)
>  	platform_device_unregister(neo1973_snd_device);
>  }
>  
> +
> +MODULE_DEVICE_TABLE(i2c, lm4857_i2c_id);
> +

No, this shouldn't be added - this driver should not be loaded based on
the presence of the LM4857 since it is specific to the OpenMoko Neo1973
system.  This is largely because the LM4857 support is a hideous bodge
which should get fixed at some point (it ought to be in a separate
driver in the CODECs directory rather than part of the OpenMoko machine
driver).

> --- linux-2.6.36-rc4-git3/sound/soc/codecs/tlv320dac33.c	2010-09-17 12:30:31.000000000 +0300
> +++ b/sound/soc/codecs/tlv320dac33.c	2010-09-17 12:39:07.801200070 +0300
> @@ -1712,6 +1712,9 @@ static int __init dac33_module_init(void
>  	}
>  	return 0;
>  }
> +
> +MODULE_DEVICE_TABLE(i2c, tlv320dac33_i2c_id);
> +

This change is OK but should be moved next to the actual table.