[PATCH 08/10] pinctrl: single: support pinconf generic

[Tony Lindgren]

* Tony Lindgren <tony at atomide.com> [121022 10:11]:
> * Haojian Zhuang <haojian.zhuang at gmail.com> [121022 03:11]:
> > On Sat, Oct 20, 2012 at 3:13 AM, Tony Lindgren <tony at atomide.com> wrote:
> > > * Haojian Zhuang <haojian.zhuang at gmail.com> [121018 02:08]:
> > >> Add pinconf generic support with POWER SOURCE, BIAS PULL.
> > > ...
> > >
> > >> +     case PIN_CONFIG_POWER_SOURCE:
> > >> +             if (pcs->psmask == PCS_OFF_DISABLED
> > >> +                     || pcs->psshift == PCS_OFF_DISABLED)
> > >> +                     return -ENOTSUPP;
> > >> +             data &= pcs->psmask;
> > >> +             data = data >> pcs->psshift;
> > >> +             *config = data;
> > >> +             return 0;
> > >> +             break;
> > >
> > > Hmm, only slightly related to this patch, mostly a generic
> > > question to others: Do others have any mux registers with
> > > status bits for things like PIN_CONFIG_POWER_SOURCE?
> > >
> > > I could use PIN_CONFIG_POWER_SOURCE for controlling the PBIAS
> > > for omap MMC. But there's also a status bit that needs to be
> > > checked for that. I think there was some other similar mux
> > > register for USB PHY that has a status register.
> > >
> > > So I'm wondering should the checking for status bit be handled
> > > in the pinctrl consume driver? Or should we have some bindings
> > > for that?
> > >
> > 
> > Do you mean that the status register only exists in USB PHY controller or
> > MMC controller?
> 
> The status register is in the MMC PBIAS register that is mux
> related otherwise. From OMAP4470_ES1.0_PUBLIC_TRM_vE.pdf,
> Table 19-599. CONTROL_PBIASLITE:
> 
> Bits
> 26	MMC1_PWDNZ
> 25	MMC1_PBIASLITE_HIZ_MODE
> 24	MMC1_PBIASLITE_SUPPLY_HI_OUT
> 23	MMC1_PBIASLITE_VMODE_ERROR	then this bit needs to clear..
> 22	MMC1_PBIASLITE_PWRDNZ
> 21	MMC1_PBIASLITE_VMODE		..after VMODE bit is set to 3V
>  
> > If so, could we use regulator framework in USB PHY or MMC driver?
> 
> Yes we could use regulator framework for that that. Or just read the
> status in the MMC driver for that bit if nobody else has mixed
> mux-regulator needs like this.
> 
> The sequence is MMC specific, so from that point of view it would
> make sense to have the logic in the MMC driver.

Well it turns out the VMODE_ERROR bit is not just for VMODE, it's a
comparator that can also triggers for the other invalid states for
CONTROL_PBIASLITE pinconf register. So hiding VMODE_ERROR into a
regulator would be wrong. For now, VMODE best handled using
PIN_CONFIG_POWER_SOURCE and let the MMC driver do the checking
using the pinconf API.

Regards,

Tony