Re: MMC: Make the configuration memory resource optional

[Ian Molton <ian@mnementh.co.uk>]

Mark Brown wrote:

Hi Mark,

> Looking at the original patch I'm not sure exactly why it runs into
> clock API issues so I'm not sure if this is a relevant concern or not
> here

The problem for tmio-mmc is that its an MFD driver. The idea behind the 
MFD framework was to allow drivers that work on similar hardware to work 
both independantly and as part of a MFD, which often have extra core logic.

Using the clock API is problematic, then, because for MFD based users of 
tmio-mmc, the clock API isnt useable, but for non-MFD users, it is.

tmio-mmc has (currently) some code in it that uses a second IO range to 
control both clocks and power on the toshiba family of MFDs. This 
ideally would be replaced by the clock API and a bunch of power control 
callbacks that would abstract it out completely from the tmio-mmc driver 
and into either platform or MFD core code depending on what the user of 
the driver is.

IOW, using the clock API will fix it for everyone.

 > bI'd kind of expect an impact on the
> SoCs from addressing it due to the way the clock API functions are
> currently provided.

Quite.

-Ian