[PATCH v2] mmc: sdhci: restore behavior when setting VDD via external regulator

[Jisheng Zhang]

Dear Ludovic,

On Fri, 11 Dec 2015 18:06:17 +0100 Ludovic Desroches wrote:

> On Fri, Dec 11, 2015 at 03:48:04PM +0100, Ulf Hansson wrote:
> > + Ludovic (We had some discussions around this code recently as well)
> >   
> 
> Thanks Ulf.
> 
> > On 11 December 2015 at 14:36, Jisheng Zhang <jszhang at marvell.com> wrote:  
> > > After commit 52221610dd84 ("mmc: sdhci: Improve external VDD regulator
> > > support"), for the VDD is supplied via external regulators, we ignore
> > > the code to convert a VDD voltage request into one of the standard
> > > SDHCI voltage levels, then program it in the SDHCI_POWER_CONTROL. This
> > > brings two issues:
> > >
> > > 1. SDHCI_QUIRK2_CARD_ON_NEEDS_BUS_ON quirk isn't handled properly any
> > > more.
> > >
> > > 2. What's more, once SDHCI_POWER_ON bit is set, some controllers such
> > > as the sdhci-pxav3 used in marvell berlin SoCs require the voltage
> > > levels programming in the SDHCI_POWER_CONTROL register, even the VDD
> > > is supplied by external regulator.So the host in marvell berlin SoCs
> > > still works fine after the commit.  
> 
> I am not sure to understand this part. You explain that the controller
> in berlin SoC requireis the voltage level programming even if there is an
> external regulator for VDD. I agree this part, I am in the same

plus one more condition ;) -- "once SDHCI_POWER_ON bit is set", that's to say
either not touching SDHCI_POWER_CONTROL register at all or setting SDHCI_POWER_ON
bit and voltage level at the same time is fine, but the sdhci-pxav3 in berlin
case can't work if we set SDHCI_POWER_ON but don't program the voltage level
, unfortunately this is true after commit 3cbc6123a93d ("mmc: sdhci: Set
SDHCI_POWER_ON with external vmmc")

> situation with atmel controller. It is not smart to rely on the voltage
> level if we have an external regulator but it follows the sdhci specs.
> 
> That I don't understand is that you say it still works fine after this
> commit... If you need to set the voltage level in the
> SDHCI_POWER_CONTROL register, it is broken by this commit if you declare
> an external regulator.

See above, commit 52221610dd84 doesn't break the host controller, it still
works fine after commit 52221610dd84 but the combination of 52221610dd84 and
3cbc6123a93d do break the host controller.

> 
> > > However, commit 3cbc6123a93d ("mmc:
> > > sdhci: Set SDHCI_POWER_ON with external vmmc") sets the SDHCI_POWER_ON
> > > bit, this would make the host in marvell berlin SoCs won't work any
> > > more with external vmmc.
> > >
> > > This patch restores the behavior when setting VDD through external
> > > regulator by moving the call of mmc_regulator_set_ocr() to the end
> > > of sdhci_set_power() function.
> > >
> > > After this patch, the sdcard on Marvell Berlin SoC boards work again.
> > >
> > > Signed-off-by: Jisheng Zhang <jszhang at marvell.com>
> > > Fixes: 52221610dd84 ("mmc: sdhci: Improve external VDD ...")
> > > ---
> > > Since v1:
> > >  - add more details about why the sdhci-pxav3 used in marvell berlin
> > >    SoCs need this patch.
> > >
> > >  drivers/mmc/host/sdhci.c | 19 ++++++-------------
> > >  1 file changed, 6 insertions(+), 13 deletions(-)
> > >
> > > diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
> > > index b48565e..616aa90 100644
> > > --- a/drivers/mmc/host/sdhci.c
> > > +++ b/drivers/mmc/host/sdhci.c
> > > @@ -1274,19 +1274,6 @@ static void sdhci_set_power(struct sdhci_host *host, unsigned char mode,
> > >         struct mmc_host *mmc = host->mmc;
> > >         u8 pwr = 0;
> > >
> > > -       if (!IS_ERR(mmc->supply.vmmc)) {
> > > -               spin_unlock_irq(&host->lock);
> > > -               mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, vdd);
> > > -               spin_lock_irq(&host->lock);
> > > -
> > > -               if (mode != MMC_POWER_OFF)
> > > -                       sdhci_writeb(host, SDHCI_POWER_ON, SDHCI_POWER_CONTROL);
> > > -               else
> > > -                       sdhci_writeb(host, 0, SDHCI_POWER_CONTROL);
> > > -
> > > -               return;
> > > -       }
> > > -
> > >         if (mode != MMC_POWER_OFF) {
> > >                 switch (1 << vdd) {
> > >                 case MMC_VDD_165_195:
> > > @@ -1345,6 +1332,12 @@ static void sdhci_set_power(struct sdhci_host *host, unsigned char mode,
> > >                 if (host->quirks & SDHCI_QUIRK_DELAY_AFTER_POWER)
> > >                         mdelay(10);
> > >         }
> > > +
> > > +       if (!IS_ERR(mmc->supply.vmmc)) {
> > > +               spin_unlock_irq(&host->lock);
> > > +               mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, vdd);
> > > +               spin_lock_irq(&host->lock);
> > > +       }
> > >  }
> > >
> > >  /*****************************************************************************\
> > > --
> > > 2.6.3
> > >  
> > 
> > My concern with this patch is that it might fix the problem for your
> > SDHCI variant, but will break it for others.
> > I guess we can give it try, unless or until someone reports a problem.
> > 
> > Although, I would like to get Ludovic's input on this change, before I
> > decide to do anything.
> >  
> 
> I would be pleased to get this patch since it would solve one of my
> issues.
> 
> Concerning the risk to take this patch. I would say one part of this
> patch is safe, the other one maybe not.
> 
> Reading the log of commit 52221610dd84, it is not a bug fix. It was done
> in this way because it seemed logical to not set the voltage level in
> the SDHCI_POWER_CONTROL if we have an external regulator.
> 
> Moving mmc_regulator_set_ocr at the end could cause issue since it
> changes the sequence order: the regulator is configured after the
> SDHCI_POWER_CONTROL register.

hmmm, this sequence order is the same as the one before commit 52221610dd84.
IOW, the patch restores the old sequence order: the regulator is configured
after the SDHCI_POWER_CONTROL register.

Thanks for reviewing,
Jisheng