On Fri, Dec 28, 2018 at 2:09 PM Tony Lindgren <tony@xxxxxxxxxxx> wrote:
>
> * Adam Ford <aford173@xxxxxxxxx> [181228 12:58]:
> > On Fri, Dec 28, 2018 at 6:41 AM Linus Walleij <linus.walleij@xxxxxxxxxx> wrote:
> > >
> > > On Thu, Dec 27, 2018 at 7:50 PM Tony Lindgren <tony@xxxxxxxxxxx> wrote:
> > >
> > > > Looks like commit e63201f19438 ("mmc: omap_hsmmc: Delete platform data
> > > > GPIO CD and WP") caused mmcblk0 to stop being detected on boot on some
> > > > omaps. According to Aaro and Adam, mmcblk0 does get detected after
> > > > replugging the SD card. Reverting the commit fixes the issue.
> > >
> > > Which boardfiles are we talking about, so I can have a look?
> > > Or are these Device Tree devices?
> >
> > I could confirm that logicpd-torpedo-37xx-devkit was fixed by
> > reverting the GPIO code or removing the pdata quirks references to the
> > pbias seem to fix the issue. I haven't tried both together yet, nor
> > have I had a chance to spend much time to fix the pbias driver to
> > correctly disconnect the IO pins when used on an AM/DM36/37 board.
> > For the Torpedo, I know the newer versions of the bootloader set the
> > pbias and IO pins correctly now (it didn't used to do that). I am
> > hoping to look at the pbias driver this weekend if I have time.
>
> Could it be that now n900 is fixed for the cover detection and
> logicpd-torpedo-37xx-devkit has the GPIO direction configured
> wrong in the dts?
I can take a look at that. The fact that i works when i remove and
reinsert the card makes me believe it's not the CD pin or wouldn't
recognize the insertion. I am not 100% certain, so I'll run some
tests. IIRC, the CD pin is one of those special omap36 pins that
needs the IO lines disconnected before the pbias switches them.
adam
>
> Regards,
>
> Tony
