Re: [PATCH 1/2] mmc: sdhci: set bus voltage before enabling bus power

[Ludovic Desroches <ludovic.desroches@atmel.com>]

On Tue, Nov 24, 2015 at 02:56:21PM +0100, Ulf Hansson wrote:
> On 24 November 2015 at 14:12, Ludovic Desroches
> <ludovic.desroches@atmel.com> wrote:
> > On Tue, Nov 24, 2015 at 12:01:53PM +0100, Ulf Hansson wrote:
> >> On 24 November 2015 at 10:23, Ludovic Desroches
> >> <ludovic.desroches@atmel.com> wrote:
> >> > Hi Ulf,
> >> >
> >> > On Mon, Nov 09, 2015 at 05:30:26PM +0100, Ludovic Desroches wrote:
> >> >> On Mon, Nov 09, 2015 at 05:00:46PM +0100, Ulf Hansson wrote:
> >> >
> >> > [...]
> >> >
> >> >> > Now, this discussion was interesting, but I forgot what problem you
> >> >> > actually where trying to solve? :-)
> >> >>
> >> >> There is this discussion because of two things:
> >> >> - Fixing something I consider as a bug: when I have a VMMC, only
> >> >>   setting/clearing bit 0. Our controller strictly obeys to the spec and
> >> >>   check the 'SD Bus Voltage Select' field. Since we put a reserved value
> >> >>   (000), the Power On is not performed.
> >> >> - I was trying to get help to understand what is this 'SD Bus Voltage'.
> >> >>   For our controller and sdhci_set_power(), it seems to stand for VMMC.
> >> >>   For me, everything concerning bus voltage is related to VQMMC, so I was
> >> >>   disappointed.
> >> >
> >> > Do you plan to take the patch for VMMC? If yes, I will send a new patch
> >> > for the device tree (I'll only add vmmc, not vqmmc as discussed); if
> >> > not, forget these two patches.
> >> >
> >>
> >> Which patch do you refer to for "VMMC"?
> >>
> >
> > This one to not write an invalid voltage in the power control register
> > even if we have an external regulator for vmmc.
> >
> 
> As I stated earlier, according to the SDHCI spec in the section for
> the Power Control Register. Bit 0 needs to be set when communicating
> with the card as it will for
> example enable the clock.
> 

I am okay with bit 0. I don't want to change this part, it will be done
later in sdhci_set_power(). My concern is only about bit 3-1, I want to
go through the switch statement.

> I suspect if I apply your patch several sdhci variants would break,
> don't you think?

I wouldn't sign it with my blood but I don't think so. It seems they
don't care about the SD bus Voltage since they work with an unsupported
voltage.

Regards

Ludovic