[U-Boot] [PATCH] mx6: Add IOMUX_CONFIG_SION flag to all GPIO pins

[Benoît Thébaudeau]

Hi Eric,

On Tuesday, October 1, 2013 4:26:32 PM, Eric Nelson wrote:
> Hi Stefano,
> 
> On 10/01/2013 06:13 AM, Stefano Babic wrote:
> > Hi Otavio,
> >
> > On 01/10/2013 14:01, Otavio Salvador wrote:
> >>> I suggest you add in the commit message that this patch is only for
> >>> i.MX6 (if you do not plan to extend it to the other i.MXes...), so that
> >>> we can track that the same must be done also for the other SOCs.
> >>
> >> I think this is clear from the commit prefix.
> >
> > Sorry, you're right, it is fine.
> >
> >>
> >> I am under heavy load now and I cannot check their datasheet to verify
> >> if they all need this or not.
> >>
> >> If someone does not does it, I can take a look for sure.
> >
> > I look at MX51 / MX53 manuals. They are different as i.MX6, and in both
> > cases I do not read the requirement to set the SION bit, as the i.MX6
> > needs. The fix is then only for i.MX6, and we do not need to extend it,
> > at least according to documentation. After this check, I think that your
> > patch is complete and I will push it for the release.
> >
> 
> I'm not sure where you're seeing this in the RM, but in
> order to read the pad state when not configured as an input,
> the SION bit needs to be set in the pad mux register on i.MX51/53.
> 
> FWIW, this also works when not muxed as a GPIO, which is really
> handy for verifying that a PWM is toggling, et cetera.

I have just tested that on i.MX51 and i.MX35, and I confirm that GPIO.PSR does
not report the correct pin level in GPIO output mode if SION is not set. The
i.MX25 is also very close to the i.MX35 in terms of IPs, so it would probably
also give the same results. So SION seems to be required in that case on all
i.MXes, even if not all RMs mention this in the GPIO chapter.

Best regards,
Benoît