[RFC PATCH v2 2/2] ARM: imx: Add mx5 cpuidle implmentation

[Rob Lee]

On 4 January 2012 23:55, Mark Brown <broonie at opensource.wolfsonmicro.com> wrote:
> On Wed, Jan 04, 2012 at 05:35:39PM -0600, Rob Lee wrote:
>> On 22 December 2011 11:50, Mark Brown
>> > On Wed, Dec 14, 2011 at 01:02:06AM -0600, Robert Lee wrote:
>
>> >> +     clk_enable(&gpc_dvfs_clk);
>
>> > Should these enables be in the cpuidle code?  The device appears to have
>> > been working fine without them thus far...  Alternatively, if they
>> > should be on anyway does this need to be split out and sent as a bug
>> > fix?
>
>> This clock is used by the existing pm_suspend code for i.MX51 and
>> other future code being worked on.  Since it uses extremely minimal
>> power and is required to be enabled during low power modes, it seemed
>> cleanest to just enable it during clock init.  But I forgot to remove
>> it from it's enabling from i.MX51 pm_suspend code so I can do that for
>> v3.
>
> Sounds like it's worth splitting out and getting it merged as quickly as
> possible then?  It wasn't the code I was querying, it was the way it is
> being merged.

Sure, I can split this portion out as a separate patch.  I'd prefer to
keep it as part of this patch series though as the existing usage and
implementation of this clock works ok, it's just not the best
implementation once another driver needs to use this clock.  And it
may raise concern if implemented by itself with a pressing reason
being shown.  At least that's my thoughts.  I'm fairly new to the
community so please tolerate anything dumb I say and help understand
the ways of the force.