Re: [PATCH v3] [MFD] i2c-htcpld: Add the HTCPLD driver

[Cory Maccarrone <darkstar6262-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>]

On Tue, Jan 19, 2010 at 4:14 AM, Jean Delvare <khali-PUYAD+kWke1g9hUCZPvPmw@public.gmane.org> wrote:
> On Tue, 19 Jan 2010 13:09:48 +0100, Samuel Ortiz wrote:
>> Hi Jean,
>>
>> On Tue, Jan 19, 2010 at 12:48:12PM +0100, Jean Delvare wrote:
>> > Hi Samuel,
>> >
>> > On Tue, 19 Jan 2010 12:09:17 +0100, Samuel Ortiz wrote:
>> > > 2) Kconfig: we need to depend on I2C=y, to avoid the problematic case of
>> > > having I2C_CORE=m and HTC_I2CPLD=y
>> >
>> > ??
>> >
>> > There is no config option named I2C_CORE. The main option for I2C is,
>> > well, I2C,
>> Sorry, I meant I2C, not I2C_CORE.
>>
>>
>> > and if you depend on I2C, then you can't be built into the
>> > kernel if I2C is built as a module.
>> I have:
>>
>> drivers/mfd/Kconfig:
>> config HTC_I2CPLD
>>       bool "HTC I2C PLD chip support"
>>       depends on I2C
>>
>> Then make menuconfig allows me to build a .config with
>>
>> CONFIG_I2C=m and CONFIG_HTC_I2CPLD=y
>>
>> and that obviously fails to build. If I'm a bool depending on I2C, it has to
>> be a built-in I2C.
>
> Ah, my bad, I totally missed that HTC_I2CPLD was a bool. Please just
> ignore me, sorry for the noise. You are correct, it should depend on
> I2C=y.
>

All these changes sound fine to me, thanks for doing that!  Sorry for
the compile mess, I'll keep all this in mind for later work.

- Cory