On Wed, Sep 13, 2023 at 2:39 PM Bartosz Golaszewski <brgl@xxxxxxxx> wrote:
> On Wed, Sep 13, 2023 at 2:24 PM Linus Walleij <linus.walleij@xxxxxxxxxx> wrote:
> >
> > On Wed, Sep 13, 2023 at 1:50 PM Bartosz Golaszewski <brgl@xxxxxxxx> wrote:
> >
> > > From: Bartosz Golaszewski <bartosz.golaszewski@xxxxxxxxxx>
> > >
> > > We have new, less cumbersome and clearer interfaces for controlling GPIO
> > > polarity. Use them in the MMC code.
> > >
> > > Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@xxxxxxxxxx>
> >
> > I like the looks of the code better, obviously but this looks like this for
> > a reason unfortunately.
> >
> > See the following from
> > Documentation/devicetree/bindings/mmc/mmc-controller.yaml:
> >
> > # CD and WP lines can be implemented on the hardware in one of two
> > # ways: as GPIOs, specified in cd-gpios and wp-gpios properties, or
> > # as dedicated pins. Polarity of dedicated pins can be specified,
> > # using *-inverted properties. GPIO polarity can also be specified
> > # using the GPIO_ACTIVE_LOW flag. This creates an ambiguity in the
> > # latter case. We choose to use the XOR logic for GPIO CD and WP
> > # lines. This means, the two properties are "superimposed," for
> > # example leaving the GPIO_ACTIVE_LOW flag clear and specifying the
> > # respective *-inverted property property results in a
> > # double-inversion and actually means the "normal" line polarity is
> > # in effect.
> >
>
> I hate it, thanks. :)
>
> > Will you still provide the desired "double inversion" after this patch?
> >
>
> Not in the current form. Would it work to go:
>
> if (override_active_level) {
> if (!(host->caps2 & MMC_CAP2_CD_ACTIVE_HIGH))
> gpiod_set_active_high(desc);
> else
> gpiod_set_active_low(desc);
> } else {
> if (host->caps2 & MMC_CAP2_CD_ACTIVE_HIGH)
> gpiod_set_active_high(desc);
> else
> gpiod_set_active_low(desc);
> }
>
> ?
I *think* so but my boolean parser i known to be flawed since I have
screwed up double inversions repeatedly over the years, so it should
not be trusted at all.
> Alternatively we could reimplement the toggle semantics locally in a
> helper function in order to get rid of it from GPIOLIB.
I don't know about that, the flag is inside gpio_desc so we cannot
access it (struct is private to gpiolib...)
Yours,
Linus Walleij
