Re: [PATCH] ASoC: dapm: Add support for multi register mux

[Lars-Peter Clausen <lars@metafoo.de>]

On 04/04/2014 09:34 AM, Arun Shamanna Lakshmi wrote:
>
>> -----Original Message-----
>> From: Lars-Peter Clausen [mailto:lars@metafoo.de]
>> Sent: Friday, April 04, 2014 12:32 AM
>> To: Arun Shamanna Lakshmi
>> Cc: lgirdwood@gmail.com; broonie@kernel.org;
>> swarren@wwwdotorg.org; perex@perex.cz; tiwai@suse.de; alsa-
>> devel@alsa-project.org; linux-kernel@vger.kernel.org; Songhee Baek
>> Subject: Re: [PATCH] ASoC: dapm: Add support for multi register mux
>>
>> On 04/03/2014 10:11 PM, Arun Shamanna Lakshmi wrote:
>> [...]
>>>> Here as well, default for bit_pos should be 0.
>>>
>>> This means when 'None' of the options are selected, by default, it
>>> enumerates to 0. Since we are using __ffs, BIT(0) of Register-0 also
>>> enumerates to 0. That's the reason why I used just ffs in the first place.
>>> Let me know your opinion. My value table looks like below.
>>>
>>> #define MUX_VALUE(npart, nbit)	(nbit + 32 * npart)
>>> static const int mux_values[] = {
>>> 	0,
>>> 	MUX_VALUE(0, 0),
>>> 	.
>>> 	.
>>> 	.
>>> 	MUX_VALUE(0, 31),
>>> 	/* above inputs are for part0 mux */
>>> 	MUX_VALUE(1, 0),
>>> 	.
>>> 	.
>>> 	.
>>> 	MUX_VALUE(1, 31),
>>> 	/* above inputs are for part1 mux */
>>> 	MUX_VALUE(2, 0),
>>> 	.
>>> 	.
>>> 	.
>>> 	MUX_VALUE(2, 31),
>>> 	/* above inputs are for part2 mux */
>>> };
>>
>> Ok, so having none of the input selected should be a valid user selectable
>> option?
>
> Yes. If 'None' is selected, it goes and clears the register. So, can we have ffs( )
> instead of __ffs( ) ? It would fix this case.

Yes, but you need to make sure to handle it also correctly in the put 
handler, since all of the registers need to be written to 0 in that case.