[PATCH] mmc: mmci: Improve runtime PM support

[ulf.hansson@stericsson.com (Ulf Hansson)]

Russell King - ARM Linux wrote:
> On Mon, Oct 24, 2011 at 11:54:00AM +0200, Linus Walleij wrote:
>> On Mon, Oct 24, 2011 at 11:36 AM, Russell King - ARM Linux
>> <linux@arm.linux.org.uk> wrote:
>>> On Sun, Oct 23, 2011 at 02:31:39AM +0200, Sebastian Rasmussen wrote:
>>>> I guess the patch would appeal more to Russell if mmci_runtime_suspend()
>>>> only cleared MCIMask0/SDI_MASK0 and MCIClock/SDI_CLKCR and left
>>>> MCIPower/SDI_PWR unchanged. It may be the case that the signal direction
>>>> bits need to be cleared for the ST-Ericsson PL180, but I haven't yet verified
>>>> this on my Snowball dev board yet.
>>> There's also the issue that the specs call for the clock to run after
>>> a command has completed for a certain number of cycles, and that the
>>> clock must continue to run until the card reports not-busy after a
>>> programming or erase cycle has completed - that can be long after the
>>> previous command has 'completed'.
>> It's 8 cycles on MCLK required.
> 
> _Plus_ keeping the clock running while the card is signalling busy.
> 
> If you don't clock the card while its signalling busy, it will never
> go non-busy (the data line becomes frozen.)
> 

Good point!

Is the aggressive clk gating feature in the mmc framework taking this 
into account as well?

Potentially the framework could do a get_sync/put (in claim/release 
host) to prevent hosts from being runtime disabled. Similar how the 
enable/disable mechanism is working.