Re: [PATCH v2] PCI: dra7xx: Fix reset behaviour

[Pali Rohár]

On Friday 25 June 2021 01:18:43 Linus Walleij wrote:
> On Wed, Jun 23, 2021 at 12:23 AM Pali Rohár <pali@xxxxxxxxxx> wrote:
> 
> > Lorenzo asked a good question how GPIO drives PERST#. And maybe it would
> > be a good idea to unify all pci controller drivers to use same GPIO
> > value for asserting PERST# pin. If it is possible. As we can see it is a
> > big mess.
> >
> > Personally I would like to a see two helper functions like
> >
> >   void pcie_assert_perst(struct gpio_desc *gpio);
> >   void pcie_deassert_perst(struct gpio_desc *gpio);
> >
> > which pci controller driver will use and we will not more handle active
> > high / low state or polarity inversion and meditate if gpio set to zero
> > means assert or de-assert.
> 
> GPIO descriptors (as are used in this driver) are supposed to hide
> and encapsulate polarity inversion so:
> 
> gpiod_set_value(gpiod, 1) == assert the line
> gpiod_set_value(gpiod, 0) == de-assert the line

Problem is that some pci controller drivers (e.g. pci-j721e.c or
pcie-rockchip-host.c) expects that gpiod_set_value_cansleep(gpiod, 1)
de-asserts the line and it is already used in this way.

Which is opposite of the behavior which you wrote above.

> Whether the line is asserted by physically driving the line low or
> high should not be a concern, that is handled in the machine
> description, we support OF, ACPI and even board files to
> define this.
> 
> I would use gpiod_set_value() directly as above and maybe
> add some comments explaining what is going on and that
> the resulting polarity inversion is handled inside gpiolib.
> 
> Because of common misunderstandings we have pondered to just
> search/replace the last argument of gpiod_set_value() from
> an (int value) to a (bool asserted) to make things clear.
> I just never get around to do that.

I would suggest to define enum/macro with word ASSERT and DEASSERT in
its name instead of just true/false boolean or 0/1 int.

In case of this PERST# misunderstanding, having assert/deassert in name
should really help.

> 
> Yours,
> Linus Walleij