Re: [PATCH 4/4] mmc: pwrseq-simple: add disable-post-power-on option

[Rob Herring <robh@kernel.org>]

On Fri, Dec 02, 2016 at 01:06:47AM -0800, Matt Ranostay wrote:
> On Fri, Dec 2, 2016 at 12:28 AM, Ulf Hansson <ulf.hansson@linaro.org> wrote:
> > On 2 December 2016 at 08:22, Matt Ranostay <matt@ranostay.consulting> wrote:
> >>
> >>
> >>> On Dec 1, 2016, at 23:13, Ulf Hansson <ulf.hansson@linaro.org> wrote:
> >>>
> >>>> On 2 December 2016 at 07:17, Matt Ranostay <matt@ranostay.consulting> wrote:
> >>>> Add optional device-post-power-on parameters to disable post_power_on
> >>>> function from being called since this breaks some device.
> >>>>
> >>>> Cc: Tony Lindgren <tony@atomide.com>
> >>>> Cc: Ulf Hansson <ulf.hansson@linaro.org>
> >>>> Signed-off-by: Matt Ranostay <matt@ranostay.consulting>
> >>>> ---
> >>>> Documentation/devicetree/bindings/mmc/mmc-pwrseq-simple.txt | 2 ++
> >>>> drivers/mmc/core/pwrseq_simple.c                            | 7 +++++++
> >>>> 2 files changed, 9 insertions(+)
> >>>>
> >>>> diff --git a/Documentation/devicetree/bindings/mmc/mmc-pwrseq-simple.txt b/Documentation/devicetree/bindings/mmc/mmc-pwrseq-simple.txt
> >>>> index bea306d772d1..09fa153f743e 100644
> >>>> --- a/Documentation/devicetree/bindings/mmc/mmc-pwrseq-simple.txt
> >>>> +++ b/Documentation/devicetree/bindings/mmc/mmc-pwrseq-simple.txt
> >>>> @@ -24,6 +24,8 @@ Optional properties:
> >>>>        de-asserting the reset-gpios (if any)
> >>>> - invert-off-state: Invert the power down state for the reset-gpios (if any)
> >>>>        and pwrdn-gpios (if any)
> >>>> +- disable-post-power-on : Avoid post_power_on function from being called since
> >>>> +       this breaks some devices
> >>>
> >>> This is a bit weird. I would like to avoid this, if possible.
> >>>
> >>> Perhaps if you simply describe the sequence in detail for your device
> >>> we can figure out the best option together.
> >>
> >> Yeah I know it is a bit weird but we need to keep that reset pin high for at least some time after pre power on.   So any case it would be another property
> >
> > This went offlist, please resend.
> >
> > Moreover, please try to describe the sequence you need in detail
> > according to your HW spec.
> >
> 
> Yeah oops....
> 
> So basically we need to drive the reset and powerdown lines high with
> a 300 milliseconds delay between both...
> can't have the reset line low with post power on (pretty sure but
> would need a delay anyway), and we need both reset + powerdown line
> set low on powerdown.
> 
> So the power down sequence would need to be reversed for this
> application in pwrseq-simple.

This sounds like you need a device specific sequence to me. Otherwise, 
write a language to describe any power control waveforms rather than 
trying to bolt on more and more properties. (Don't really go write a 
language.) 

Rob